A Guide for Monitoring and Evaluating Child Health ... - CORE Group

A Guide for Monitoring and EvaluatingChild Health ProgramsAnastasia J. GageDisha AliChiho SuzukiEvaluationThis guide was made possible by support from the U.S. Agency for International Development (USAID)under terms of Cooperative Agreement GPO-A-00-03-000003-00. The author's views expressed inthis publication do not necessarily reflect the views of USAID or the United States Government.September 2005MS-05-15Printed on rececycled paper

ACKNKNOWLEDGEMENTWLEDGEMENTSThis Guide has grown out of the collective experience of many individuals and organizations over thepast few decades in monitoring and evaluating child health programs at the national level in developingcountries. The effort, coordinated by the United States Agency for International Development (USAID),World Health Organization Headquarters (WHO-HQ), and MEASURE Evaluation, was co-financedby the USAID Office of Health, Infectious Diseases and Nutrition (HIDN) and WHO. We areparticularly grateful to Al Bartlett and Norma Wilson of USAID/HIDN, Thierry Lambrechts andcolleagues of the WHO Department of Child and Adolescent Health and Development, and MariaFrancisco, who sustained this activity with enthusiastic and unwavering support and sound technicaladvice.The process has involved repeated consultation with monitoring and evaluation specialists at WHO/AFRO and the United States Centers for Disease Control and Prevention (CDC). We are particularlygrateful to Dr. Wondi Alemu, who led the significant amount of work that went on with partners inCDC, WHO-HQ and Ministries of Health in the African region to develop the set of indicators formonitoring and evaluating integrated disease surveillance and response, which are described in ChapterFive of the guide. The United Nations Children’s Fund (UNICEF) and the World Bank providedsubstantial input to the development of the guide.Technical input has come from numerous colleagues and a number of institutions, including WHO-HQ, CDC, Academy for Educational Development/FANTA and Change, BASIC Support forInstitutionalizing Child Survival II Project, the Environmental Health Project, John Snow InternationalResearch and Training Institute, ORC Macro International/Child Survival Technical Support andMEASURE DHS, Project Hope/The Core Group, Save the Children, and MEASURE Evaluation.A technical advisory group meeting was held in May 2003 at which a draft version of this Guide wasextensively reviewed. This meeting was attended by technical specialists from a range of institutions.Following this meeting, the draft guide was revised and finalized.The authors are especially grateful to the following colleagues (listed below in alphabetical order) whocontributed to this volume in various ways through attending meetings, preparing text, reviewing draftsof specific chapters, and answering requests for information:Mounkaila AbdouJacob AdetunjiJuan Carlos AlegreSharon Arscott-MillsStacey BallouSandra BertoliEd BosMatthew CarneyMisun ChoiEunyong ChungStella ChungongBruce CogillSiân CurtisRebecca FieldsMaria FranciscoMary FreyderNancy FronczakAmanda GibbonsAnja GiphartStephen HadlerPaige HarriganEckhard KleinauAnn LaFondThierry LambrechtsMargaret LamunuJoy LawnRobert MagnaniJohn MasonAcknowledgementsiii

Jay McAuliffeSubhi MehdiRoy MillerEric MintzJohn NovakPeter NsubugaMichel PacquéHelen N. PerryBeth PlowmanSandra RemancusAmanda RoseLa Rue SeimsKavita SinghAdam SloteCynthia StantonMary Ellen StantonRobert SteinglassPatricia StephensonMurray TrostleSteve WallTessa WardlawNorma WilsonProfessor D.L. WoodsJelka ZupanRecommended citation:Gage, Anastasia J., Disha Ali, and Chiho Suzuki. (2005). A Guide for Monitoring and EvaluatingChild Health Programs. MEASURE Evaluation. Carolina Population Center, University of NorthCarolina at Chapel Hill.ivA Guide for Monitoring and Evaluating Child Health Programs

TABLEOF CONTENTONTENTSAcknowledgements ............................................................................................................................. iiiList of Figures and Tables ................................................................................................................... viiSummary List of Indicators ................................................................................................................. ixAcronyms ................................................................................................................................... xvChapter 1. Overview ............................................................................................................................1A. Rationale for the Guide .............................................................................................................1A.1 Objectives of the Guide ....................................................................................................3A.2 Intended Audience ............................................................................................................4A.3 Organization of the Guide ................................................................................................4B. Indicators for Program Monitoring and Evaluation ..................................................................5B.1 Program Components - Inputs, Processes, Outputs, and Outcomes ................................5B.2 Defining Monitoring and Evaluation ................................................................................6B.3 Program-based and Population-based Measures ...............................................................8B.4 Monitoring and Evaluation during Phases of the Program Cycle.....................................9B.5 Value of a Conceptual Framework .....................................................................................9B.6 Indicators .........................................................................................................................13B.7 How to Select Indicators .................................................................................................14B.8 Data Sources ....................................................................................................................15B.8.1 Household and community level data ....................................................................15B.8.2 Health systems data ................................................................................................17B.8.3 Qualitative data ......................................................................................................18B.9 The Use of Indicators at Different Levels .......................................................................19Chapter 2. Prevention of Mother-to-Child Transmission of HIV ....................................................24Chapter 3. Newborn Health ...............................................................................................................42Chapter 4. Immunization .................................................................................................................112Chapter 5. Integrated Disease Surveillance and Response ...............................................................160Chapter 6. Integrated Management of Childhood Illness: Improved Health Worker Skills ...........206Chapter 7. Diarrhea, Acute Respiratory Infection, and Fever ..........................................................240Chapter 8. Growth Monitoring and Nutrition ................................................................................302Chapter 9. Mortality .........................................................................................................................370Table of Contentsv

LISIST OF FIGURESAND TABLESFigure 1.1. Program phases and priority monitoring and evaluation activities ..................................10Figure 1.2. Conceptual framework for monitoring and evaluating child health programs ................12Table 3.1. WHO essential newborn care package ............................................................................44Figure 3.1. Maternal & newborn health linkages ..............................................................................45Table 3.2. Essential newborn care intervention sub-package ...........................................................55Table 3.3. WHO recommended tetanus toxoid immunization schedule for women ofchildbearing age and pregnant women without previous exposure to TT, Td, or DTP ..66Table 3.4. Essential newborn care intervention sub-package ...........................................................80Table 3.5. Potential causes of death for specific age and birth weight categories ...........................102Table 4.1. Illustrative inputs, processes, outputs, and outcomes for monitoring immunizationprograms.........................................................................................................................114Table 4.2. Tetanus toxoid immunization schedule for women of childbearing age and pregnantwomen without previous exposure to tetanus toxoid-containing vaccines(TT, Td, or DTP)...........................................................................................................120Figure 5.1. Information flow in an integrated disease surveillance system (IDS) ............................163Figure 5.2. Steps for establishing health facility thresholds .............................................................176Figure 5.3. The role of laboratories in a well-functioning surveillance system ................................182Table 5.3. Recommended responses to confirmed outbreaks for selected diseases that are ofrelevance to child health programs.................................................................................186Figure 6.1. Outline of the integrated management of childhood illness (IMCI) impact model .....207Table 6.1. Illustrative processes, outputs, and outcomes associated with improving healthworker skills ...................................................................................................................208Table 7.1. List of indicators for the household and community component of IMCI ...................242Table 7.2. RBM core indicators ......................................................................................................244Figure 8.1. Conceptual framework for the causes of malnutrition in society...................................304Figure 8.2. Conceptual framework for monitoring and evaluating child nutrition programs .........305Table 8.1. PAHO/WHO guiding principles for complementary feeding of the breastfed child ...312Table 8.2. Components and scoring of a young child feeding practices score for children aged6-23 months ...................................................................................................................333List of Figures and Tablesvii

SUMMARUMMARY LISIST OF INDICNDICATORSThe indicators in this guide are organized into eight chapters. Chapters two toeight describe indicators that are relevant to specific programmatic areas of childhealth. Chapter IX presents mortality indicators. Essential core indicators that arerelevant for monitoring and evaluation in multiple programmatic areas have beencross-referenced. For example, some hygiene indicators are cross-referenced in thenutrition and diarrhea/ARI/fever sections. The indicators appearing within eachchapter are detailed below.••••••••••••••••Summary List of Indicatorsix

•••••••••••••••••••••••••••••••••xA Guide for Monitoring and Evaluating Child Health Programs

•••••••••••••••••••••••••Summary List of Indicatorsxi

•••••••••••••••••••••••••••xiiA Guide for Monitoring and Evaluating Child Health Programs

•••••••••••••••Summary List of Indicatorsxiii

ACRCRONYMSACC/SCNAEFIAFPAFRAMROANCAPHAPIARIARTARVAZTBASICSBCCBCIBCGBFHIBHRC-IMCICATCHCFRCBDCDCCDDCHWCOREDDDHSDTPEBREHEHPENDENMREOCEPIFANTAFICFPGAVIGDPHAARTHIFAdministrative Committee on Coordination/StandingCommittee on NutritionAdverse Events Following ImmunizationAcute Flaccid ParalysisBureau of AfricaWHO Latin American Regional OfficeAntenatal ClinicAnte-Partum HemorrhageAnnual Parasite IndexAcute Respiratory InfectionAntiretroviral TherapyAntiretroviral (drugs)Azidodeoxythymidine (Retrovir)Basic Support for Institutionalizing Child Survival IIBehavior Change CommunicationBehavior Change InterventionsBacille Calmette-Guerin (vaccine)Baby Friendly Hospitals InitiativeBureau of Humanitarian ResponseCommunity Integrated Management of Childhood IllnessCore Assessment Tool on Child HealthCase Fatality RateCommunity-Based DistributionCenters for Disease Control and PreventionControl of Diarrheal DiseaseCommunity Health WorkerChild Survival Collaborations and Resources GroupDiarrheal DiseaseDemographic and Health SurveyDiphtheria, Tetanus, Pertussis (vaccine)Exclusive Breastfeeding RateEnvironmental HealthEnvironmental Health ProjectEarly Neonatal DeathEarly Neonatal Mortality RateEmergency Obstetric CareExpanded Program on Immunization (WHO)Food and Nutrition Technical AssistanceFully Immunized ChildFamily PlanningGlobal Alliance for Vaccines and ImmunizationGross Domestic ProductHighly Active Antiretroviral TherapyHygiene Improvement FrameworkAcronymsxv

HFAHFSHHIQATHIBHIFHISHIVHMISHMNHPLCIAWGICCIDDICDSICHSIDDIECIMCIIPTITMITNIUIUGRIVACGJMPKAPKPCLBWlcdLFDRLMISLNMRLQASMEASUREMEWGMCHMDGM&Emg/lMICSMISMMRMOHMTCTmU/lNCHSNIDSNGOxviHealth Facility AssessmentHealth Facility SurveyHousehold Hygiene Improvement Quantitative Assessment ToolHaemophilus Influenza Type B (vaccine)Health Improvement FrameworkHealth Information SurveyHuman Immunodeficiency VirusHealth Management Information SystemsHealth Metrics NetworkHigh-Performance Liquid ChromatographyInter-Agency Working GroupInternational Council for Control of IDDIntegrated Child Development ServicesIntegrated Child Health SurveyIodine Deficiency DisordersInformation, Education and CommunicationIntegrated Management of Childhood IllnessIntermittent Preventative TreatmentInsecticide-Treated MaterialsInsecticide-Treated NetInternational UnitsIntrauterine Growth RetardationInternational Vitamin A Consultative GroupJoint Monitoring ProgrammeKnowledge, Attitude, and PracticeKnowledge, Practice, CoverageLow Birth WeightLiters Per Capita Use Per Day (lcd)Late Fetal Death RateLogistics Management Information SystemLate Neonatal Mortality RateLot Quality AssuranceMonitoring and Evaluation to Assess and Use ResultsMonitoring and Evaluation Working GroupMaternal and Child HealthMillennium Development GoalMonitoring and EvaluationMilligrams Per LiterMultiple Indicator Cluster SurveyManagement Information SystemMaternal Mortality RateMinistry of HealthMaternal to Child TransmissionMilliunits Per LiterNational Center for Health StatisticsNational Immunization DaysNongovernmental OrganizationA Guide for Monitoring and Evaluating Child Health Programs

NMRNRLNTNTMROSDOPVORSORTPAHOpHPHCPMRPMTCTPNDAPpmPVCPVOPVPQAQIQRAMOSRBMRBPRHSRTHSDSESSDPSNIDSPSPASTHSTITDTSHTTU5MRµgUNUNAIDSUNDPUNFPAUNGASSUNICEFUNICEF/ESARUSAIDAcronymsNeonatal Mortality RateNational Reference LibraryNeonatal TetanusNeonatal Tetanus Mortality RateOffice of Sustainable DevelopmentOral Polio VaccineOral Rehydration SaltOral Rehydration TherapyPan American Health OrganizationPotential of HydrogenPrimary Health CarePerinatal Mortality RatePrevention of Mother to Child Transmission (of HIV)Perinatal Death AuditsParts Per MillionOffice of Private and Voluntary CooperationPrivate Voluntary OrganizationPredictive Value PositiveQuality AssuranceQuick Investigation of QualityReproductive Age Mortality SurveyRoll Back MalariaRetinol Binding ProteinRecommended Home FluidRoad to HealthStandard DeviationSocioeconomic StatusService Delivery PointSub-National Immunization DaySulfadoxine-PyrimethamineService Provision AssessmentSoil-Transmitted HelminthsSexually Transmitted InfectionTetanus-DiptheriaThyroid Stimulating HormoneTetanus Toxoid (vaccine)Under-Five Mortality RateMicrograms (millionth of a gram)United NationsJoint United Nations Program on HIV/AIDSUnited Nations Development ProgrammeUnited Nations Population FundUnited Nations General Assembly Special SessionUnited Nations Children’s FundUnited Nations Children’s Fund/Eastern & Southern AfricaRegional OfficeUnited States Agency for International Developmentxvii

UTIVADVADDVCTVEVIPVVMWHOWHO/AFROWFSWHSWSSCCUterine Tract InfectionVitamin A DeficiencyVitamin A Deficiency DisorderVoluntary Counseling and TestingVaccine EfficiencyVentilated Improved LatrineVaccine Vial MonitorsWorld Health OrganizationWorld Health Organization/Regional Office for AfricaWorld Fertility SurveyWorld Health SurveyWater Supply and Sanitation Collaborative CouncilxviiiA Guide for Monitoring and Evaluating Child Health Programs

1HAPTER 1. OCHAPTER1. OVERVERVIEWVIEWA. Rationale for the GuideAnumber of recent global initiatives havetriggered renewed interest in supporting andstrengthening the monitoring and evaluation ofchild health programs. The MillenniumDevelopment Goals (MDGs) are among the mostprominent and provide a commonly acceptedframework, benchmarks, and indicators formeasuring development progress. Six of theMDGs (see Annex Table 1.1) are directly relevantto children and match the goals set out in “A WorldFit for Children” (United Nations, 2002). Goals4, 5, and 6 set out to reduce child mortality,improve maternal health, and combat HIV/AIDS,malaria and other diseases, respectively. TheMDGs have not only accelerated demand for data,but they have also highlighted limitations in dataavailability and quality, as significant numbers ofcountries do not have enough data to track changesin poverty, child malnutrition and HIV/AIDSprevalence. Some countries also face serious dataquality issues in measuring maternal mortality andaccess to water and sanitation. Achieving theMDGs by 2015 will require greater focus onoutcomes rather than inputs to effectively measureprogress at the national and global levels.Internationally-led efforts to monitor and evaluateprogress in reducing child mortality and morbidityalso encompass the Integrated Management ofChildhood Illness (IMCI) strategy. The strategyincludes interventions to prevent illness and reducedeaths from the most common child healthproblems and to promote child health anddevelopment. The interventions comprise threecomponents: improving health worker casemanagement skills, improving the health systemto deliver IMCI, and improving family andcommunity practices. The World HealthOrganization (WHO) Department of Child andAdolescent Health and Development hasdeveloped recommendations for monitoringIMCI at national and district levels, conductedworldwide monitoring of IMCI implementation,developed tools for national and district levelevaluation, and established milestones forworldwide monitoring of IMCI implementation(WHO, 1999). A list of priority and supplementalindicators for IMCI implementation at first levelhealth facilities and in the community wasdeveloped and agreed upon by an Inter-AgencyWorking Group on IMCI Monitoring andEvaluation for use in all monitoring and evaluationactivities to facilitate the collection of comparableinformation in different settings.The Global Alliance for Vaccines andImmunization (GAVI) was launched by itspartners in 2000 to fight declining immunizationrates and growing disparities in access to vaccinesamong the world’s poorest countries. GAVI is apublic-private partnership between governmentsin developing and industrialized countries,established and emerging vaccines manufacturers,nongovernmental organizations (NGOs), researchinstitutes, United Nations Children’s Fund(UNICEF), World Health Organization (WHO),the Bill & Melinda Gates Foundation, and theWorld Bank. GAVI’s objectives are to: improveaccess to sustainable immunization services;expand the use of all existing safe and cost-effectivevaccines, and promote delivery of otherappropriate interventions at immunizationcontacts; support the national and internationalaccelerated disease control targets for vaccinepreventablediseases; accelerate the developmentand introduction of new vaccines and technologies;accelerate research and development (R&D)efforts for vaccines needed primarily in developingcountries; and make immunization coverage acenterpiece in international development efforts.To help measure progress towards its overall goalChapter 1. Overview1

of protecting children of all nations andsocioeconomic levels against vaccine-preventablediseases, GAVI has established the followingmilestones:(1) By 2010 or sooner all countries will haveroutine immunization coverage at 90%nationally with at least 80% coverage in everydistrict.(2) By 2007, all countries with adequate deliverysystems will have introduced hepatitis Bvaccine.(3) By 2005, 50% of the poorest countries withhigh disease burdens and adequate deliverysystems will have introduced Hib vaccine.(4) By 2008, the world will be certified polio-free.(5) By 2005, the vaccine efficacy and burden ofdisease will be known for all regions forrotavirus and pneumococcal vaccine, andmechanisms will be identified to make thevaccines available to the poorest countries.International efforts in monitoring and evaluatingprograms for the prevention of HIV in infants andyoung children were spurred in large measure bythe adoption of the Declaration of Commitmenton HIV/AIDS at the United Nations GeneralAssembly Special Session on HIV/AIDS in June2001, at which governments from 185 countriescommitted themselves to a comprehensiveinternational and national effort to fight the HIV/AIDS pandemic. The Declaration established anumber of goals for the achievement of specifictargets including reductions in HIV infectionamong infants and young adults; improvementsin HIV/AIDS education, health care andtreatment; and improvements in orphan support.The Declaration of Commitment has generatedan international effort, led by the Joint UnitedNations Programme on HIV/AIDS (UNAIDS),to develop a core set of indicators for monitoringvarious aspects of national and internationalactions, national program outcomes and nationalimpact objectives (UNAIDS, 2002). Some of theglobal and national indicators for implementationof the Declaration of Commitment have a directbearing on child health programs. Ongoing effortsinclude the development of guidelines for LocalMonitoring and Evaluation of the IntegratedPrevention of Mother to Child HIV/Transmissionin Low-Income Countries (UNAIDS, 2000) and thedevelopment of a national guide to monitoringand evaluating programmes for the prevention ofHIV in infants and young children.Another important initiative that has shapedinternational monitoring and evaluating efforts isRoll Back Malaria (RBM). Roll Back Malaria isa global partnership founded in 1998 by WHO,the United Nations Development Programme(UNDP), UNICEF, and the World Bank with thegoal of halving the world’s malaria burden by 2010.The RBM partnership includes nationalgovernments, civil society and nongovernmentalorganizations, research institutions, professionalassociations, UN and development agencies,development banks, the private sector and themedia. Drawing on past work accomplished bythe WHO Regional Offices and on the morerecent efforts of the WHO Regional Offices forAfrica and for Eastern Mediterranean during theAccelerated Implementation of Malaria Controlin 1997-1998, WHO has developed a frameworkalong with indicators for monitoring the progressand evaluating the outcomes and impacts of RBM.A multi-disciplinary group on RBM monitoringand evaluation was created to propose a frameworkfor monitoring and evaluation to be endorsed byall RBM partners, to specify, as far as possible,standard methods, indicators and criteria to beused, to select a set of tools to be used to collectthe data needed for measuring global indicatorsand to propose a guide for use of these tools(WHO, 2000).Faced with the necessity to improve countrycapacity to provide and use health information andthe increasing data needs among the global donorcommunity, a new global initiative, the HealthMetrics Network (HMN), was launched in May2005. The overarching goal of the HMN is toimprove the availability and quality of population2Chapter 1

and health information in resource poor countries.It proposes to do this through three primaryfunctions: global coordination of healthinformation efforts, strengthening country healthinformation systems, and development andpromotion of priority innovations in healthinformation methods. The HMN is a network ofpartners, spearheaded initially through the effortsof the Bill and Melinda Gates Foundation andWHO, that will work with ministries of health,departments of statistics and information,international organizations, multi-lateral andbilateral agencies, foundations, academicinstitutions and civil society organizations. Theprimary focus of the HMN will be at country leveland activities will be country led. The HMN willwork to mobilize resources and technical expertiseto assist countries in their health informationsystem improvements and reform. At the sametime the HMN will work with global partners incollaborating and coordinating global monitoringand evaluation efforts in order to lessen the burdenplaced on country partners as they respond toinformation requests.This is by no means an exhaustive list of initiativesthat have renewed interest in the monitoring andevaluation of health programs. The importanceof health information as a public good has gatheredmomentum since the adoption of the MDGs.International meetings on selected health topicshave provided a forum for generatingrecommendations for outcome indicators, a casein point being the International Vitamin AConsultative Group (IVACG) Meeting held fromOctober 30 to November 2, 2000 in Annecy,France where outcome indicators were proposedfor monitoring and evaluating vitamin A programs(Wasantwisut, 2002). The process of developinginternational consensus on monitoring andevaluation frameworks, indicators, and tools hastypically involved consultations among globalpartners and has taken considerable time andeffort. As a result, many of the indicators proposedby these initiatives have achieved broadinternational consensus and have been tested andused extensively in the field. The current volumereaps the benefits of these recent initiatives, as wellas of earlier international efforts to developindicators on various aspects of child health.A.1 Objectives of the GuideThe overarching objective of this guide is toencourage program monitoring and evaluation andto improve the quality of work in the child healtharea. To this end, the guide provides acomprehensive listing of the most widely usedindicators for monitoring and evaluating childhealth programs in developing countries. Theindicators are organized using a generic conceptualframework. This framework maps the pathwaysthrough which programs achieve results (seeFigure 1.2 on page 12), and as such constitutes alogical framework for developing a monitoring andevaluation plan with appropriate indicators. Manyof the program areas covered in the guide containmore detailed frameworks that explain thepathways for program effects specific to differenttechnical intervention areas.Past monitoring and evaluation efforts havesometimes focused exclusively on child healthoutcomes or ultimate program results, with littleclarification of how programs operate to achievetheir desired results. However, this frameworkspecifies how those who design the program expectit to work to achieve results at both the programand population level. Moreover, the frameworkdraws attention to the different aspects of childhealth programs that must be workingsatisfactorily to achieve the desired end result.The specific objectives of this guide are to:• Compile indicators judged to be most usefulfor monitoring and evaluating child healthprograms.• Encourage the consistent use of standardizeddefinitions of indicators and terminologyacross the child health community.Overview3

• Serve as a central source for obtainingmeasures of process and output that can bereasonably linked to program activities.• Promote the monitoring and evaluation ofchild health programs by making indicatorsbetter known and easier to use.A.2. Intended AudienceSeveral different audiences should find this guidepertinent to their own work, including:• Directors, managers and staff of child healthprograms worldwide• Staff of international agencies dealing withchild health• Monitoring and evaluation specialists• Applied researchersA.3. Organization of the GuideChapter 1 provides an overview of the Guide andpresents generic monitoring and evaluationterminology that has been used to organize theindicators. This section describes hypotheticalphases during the life of a program whenevaluation can make a contribution; types ofevaluation; and the major sources of data on whichindicators are based. It also provides guidance onhow to select indicators for program monitoringand evaluation.The rest of the volume is organized by major areaof program intervention. These encompass theprevention of mother-to-child transmission ofHIV/AIDS, newborn health, immunization,integrated disease surveillance and response,integrated management of childhood illness at thehealth facility level, diarrhea/acute respiratoryinfection (ARI)/fever (malaria), growthmonitoring and nutrition, and mortality. Eachchapter briefly discusses measurement challengesand presents indicators for monitoring andevaluating the service delivery environment,specifically access to services and quality, and keypopulation-based outcomes for the programmaticarea under consideration.4Each indicator is described in the context ofprogram goals. The data requirements aresummarized, and reference is made to thequestionnaires or measurement tools that wouldprovide the required information for constructingthe indicator. Details on methods of calculationare given for each indicator. Where calculation ofthe indicator requires the initial computation of anumerator and a denominator, precise definitionsof these components are provided. The guidelinesfor each indicator end by highlighting theindicator’s strengths and limitations and points tobe considered when interpreting estimates of theindicator. Particular attention is paid tohighlighting factors that could distort trends inthe indicator as these may lead to incorrectconclusions being drawn on program effectiveness.We also include targets and benchmarks, wherethese have been established at the global level byinternational consensus. In some poor countries,many of these targets and benchmarks may seemfar out of reach. Even in better-off countries theremay be regions or groups that lag behind.Individual countries may set and monitor progressagainst their own internal targets and benchmarks,if they wish to do so.To the extent possible, we selected indicators thathad been field-tested, including those measuredin the Demographic and Health Surveys (DHS)and Multiple Indicator Cluster Surveys (MICS).However, some of the indicators presented in thisguide represent work in progress. In some cases,data for measuring these indicators are not yetavailable or not adequately collected throughinternational survey efforts.This manual does not address output indicatorsfor the multiple functional areas that are essentialto support program activities. These functionalareas include management, capacity building,commodities and logistics, behavior changecommunication (BCC), policy, and advocacy.Despite the importance of these topics to childhealth programs, the authors were not able tolocate a standard set of indicators that had beenChapter 1

tested and were in use at the field level to monitorand evaluate the functional areas of child healthprograms. The Child Health Technical AdvisoryGroup felt that such indicators would need to becontext-specific and proposed that a separate effortbe initiated to develop a set of guidelines that willassist organizations in identifying appropriatemeans of monitoring and evaluating thesefunctional areas within their organizations. Thecurrent volume focuses, therefore, on indicatorsmeasuring the adequacy of the service deliveryenvironment as well as population-basedoutcomes.Two topics that have not been covered by themanual are cost analysis or cost effectiveness andsustainability. These topics were excluded due tothe difficulty of locating a set of standardizedindicators on which international consensus hadbeen reached. Some emerging areas are also notcovered. These include gender, the urban poor,social mobilization, scaling-up, equity/povertyreduction, orphans and vulnerable children, andcomplex emergencies. It was felt that newindicators are not needed for some of these areasand that some issues like gender, urban povertyand equity can be addressed at the tabulation phaseby disaggregating many of the indicators asappropriate. In this regard, the STATcompilersoftware program on the MEASURE DHS Website can be utilized as one way of obtainingindicators pertaining to equity, the urban poor, ororphans.While the guide includes many of the indicatorsrecommended for monitoring and evaluatingIntegrated Management of Childhood Illness(IMCI), some of the key family practices proposedby UNICEF for improving care of children at thehousehold and community levels are notablyabsent. These include:(1) Promote children’s mental and social developmentby being responsive to the child’s needsfor care, and by stimulating the child throughtalking, playing, and other appropriate physicaland affective interactions.(2) Take steps to prevent child abuse, recognizeit has occurred, and take appropriate action.(3) Adopt and sustain appropriate behaviorregarding HIV/AIDS prevention and care forthe sick and orphans.(4) Ensure that men actively participate inproviding childcare, and that they are involvedin reproductive health initiatives.(5) Prevent and provide appropriate treatment forchild injuries.Interest in monitoring and evaluating the successof programs designed to influence these areas hasgrown. However, more work needs to be done formeasuring progress in these areas. Therefore,relevant indicators for these key family practiceshave been excluded from the document. Readersare referred to the following Web site for furtherinformation: http://www.unicef.org/programme/health/focus/community/cimci/overview.htmB. Indicators for Program Monitoring andEvaluationB.1.Program Components – Inputs, Processes,Outputs, and OutcomesAs with other public health programs, child healthprograms may be thought of as consisting of a setof components (defined below). Throughout theguide, the discussion of monitoring and evaluationand indicators is organized around thesecomponents:• Inputs refer to the human and financialresources, physical facilities, equipment,clinical guidelines, and operational policiesthat are the core ingredients of child healthprograms and enable health services to bedelivered.• Processes refer to the multiple activities thatare carried out to achieve the objectives ofchild health programs. Although a high levelof input is generally reflected in satisfactoryprogram implementation, it is theoreticallypossible to have a high level of inputs but apoorly delivered program (for example,Overview5

available resources that are poorly managedor quality that is not monitored).Conversely, there are countless real-lifeexamples around the world, where programstaff with inadequate resources strive to do thebest work they can under the circumstances.• Outputs refer to the results of these effortsat the program level. Although child healthprogram managers at the field level areinterested in national/sub-national trends inchild morbidity, nutrition, and mortality, theytend to limit the monitoring and evaluationof their own activities to program-basedmeasures, especially measures of output. Twotypes of outputs may be distinguished:• Functional outputs, which measure thenumber/quantity of activities conductedin each functional area of service delivery,such as behavior change communication,commodities and logistics, managementand supervision, training, etc.• Service outputs, which measure thequantity of services provided to theprogram’s target population, as well asthe adequacy of the service deliverysystem in terms of access, quality of care,and program image/client satisfaction.• Outcomes refer to changes measured at thepopulation level in the program’s targetpopulation, some or all of which may be theresult of a given program or intervention.Outcomes refer to specific knowledge,behaviors, or practices on the part of theintended audience – such as timely initiationof breastfeeding, increased fluids andcontinued feeding during illness, and increaseduse of oral rehydration therapy – that areclearly related to the program, can reasonablybe expected to change over the short-tointermediateterm, and that contribute to aprogram’s desired long-term goals. Outcomesalso include coverage and disease prevalence.• Impact refers to the anticipated end results of aprogram – for example, reducing diseaseincidence, improving children’s nutritional status,and reducing child morbidity and mortality.B.2. Defining Monitoring and EvaluationMonitoringMonitoring is the routine tracking of a program’sactivities by measuring on a regular, ongoing basiswhether planned activities are being carried out.Monitoring systems inform managers whetherprogram activities are being implemented accordingto plan and at what cost, how well the program isfunctioning at different levels, the extent to which aprogram’s services are being used, whether interimtargets are being met, and whether key performancemeasures are changing.The terms “monitoring” and “process evaluation” areoften used interchangeably. Process evaluationmeasures how well program activities are beingperformed. This information is sometimes collectedon a routine basis, such as through staff reports, butmay also be collected periodically in a larger-scaleprocess evaluation effort (i.e., special studies) thatmay include use of observational studies, surveys ofclients, focus groups or other qualitative methods.Process evaluation is often used to measure thequality of program implementation and to assesscoverage; it may also measure the extent to whichservices are used (Adamchak et al., 2000). Programmanagers, staff, and participants in a program tendto be the primary users of process evaluations butthis does not preclude any other type of stakeholderfrom using the findings.In this guide, we distinguish between the two bywhether the data in question are gathered routinely– “monitoring” entails the assessment of programoperational performance based upon routinelycollected information, while “process evaluation” alsoentails non-routine data collection and often lessstructured/more open-ended approaches to collectinformation on the strengths and weaknesses of theprogram. As the purpose of both monitoring andprocess evaluation is to assess program operationalperformance, readers who prefer alternativedistinctions between the two terms or use them6Chapter 1

interchangeably should not encounter majordifficulties in using the indicators presented in theguide.EvaluationIn the early days of program evaluation, and tosome extent in contemporary child healthprograms, evaluation tended to focus on outcomes– the extent to which intended population-levelchanges, as defined by the objectives of theprogram, are achieved. However, that is changingrapidly. In order to improve a program, it isnecessary to understand how well it is movingtoward its objectives so that changes can be madein the program components (Herman et al., 1987).Ignoring implementation issues limits theusefulness of findings about effective programs andis a major impediment to improving complexoperating programs or conducting policy analysis.The multiple informational needs for managingand assessing the results of child health programsrequire the use of complementary evaluationapproaches and methodologies. Two main typesof program evaluation may be distinguished:formative evaluation and summative evaluation.Formation EvaluationFormative evaluation is conducted during theplanning or re-planning stage of a program toidentify priority unmet health needs, barriers andconstraints to the use of health services, and factorsunderlying existing health problems anddisparities. A formative evaluation may include aneeds assessment to determine what are thespecific needs with regard to child health servicesand the best ways to meet those needs. A needsassessment may also be conducted to establishbaseline measurements for documenting changesin service delivery as a result of this project. Forthe purpose of the guide, we will assume that theprogram design or initiation phase has beencompleted.Summative EvaluationSummative evaluations address establishedinterventions and are used to make decisions aboutthe program being evaluated. Evaluations ofprogram efficacy are conducted when interventionsare delivered through health services in relativelyrestricted areas and assess whether, given idealcircumstances, the intervention had an effect(Bryce et al., 2004). Evaluations of programeffectiveness assess whether the interventions havean effect under “real-life” circumstances faced byhealth services. Few public health programs areimplemented in ways that allow evaluations to beentirely “efficacy” or entirely “effectiveness”(Habicht et al., 1999).An important component of summativeevaluations is establishing the level of certaintythat decision-makers need to have that anyobserved effects are in fact due to the project orprogram. Adequacy evaluations assess how well theprogram activities have met the expected objectives– whether or not the expected changes have takenplace. These may include assessments of howmany health centers have been opened, how manyORS packets or other drugs are available, how wellworkers have been trained, how many childrenused the service, what coverage has been achievedin the target population or whether health andbehavioral indicators have improved among thetarget population. Adequacy evaluations do notrequire a control group and may be cross-sectionalor longitudinal (Habicht et al., 1999).Plausibility evaluations refer to whether changesin indicators – be they service provision, utilization,coverage, or impact – are likely to be due to theintervention. Plausibility evaluations try to ruleout the influence of external or confoundingfactors and require a control group (Bryce et al.,2004). Finally, probability evaluations requirerandomization of treatment and control activitiesand aim at ensuring that there is only a smallknown probability that the differences betweenOverview7

program and control areas were due to chance.Probability evaluations are often not feasible foraddressing program effectiveness for severalreasons. Evaluators must be present at a very earlystage of the program planning cycle to design therandomization of services, communities orindividuals to intervention or control groups butare often not recruited until after the program hasbeen implemented. There may also be politicalfactors influencing the placement of the newinterventions. Additionally, the methodologyoften results in situations that are different fromreality and not useful for the decisions that are tobe made (Habicht et al., 1999).B.3. Program-based and Population-based MeasuresFor the purpose of monitoring and evaluation, itis important to distinguish between programbasedand population-based measures/data.Program-based data consist of informationavailable from program sources (e.g.,administrative records, client records, servicestatistics) or information that can be obtained fromon-site collection (e.g., observation, clientproviderinteraction, client exit interviews,simulated purchase/mystery client surveys),although routine health information systems arealso the primary source of program-based data.Also, a follow-up study of clients who attended aclinic constitutes program-based data, in that theinformation on the clients comes from programrecords. Although some program-based datacorrespond to a limited network of clinicsproviding a specialized service, “program-based”can also refer to programs that are national inscope.Program-based information is very important forunderstanding the performance of programs andthe type of output they achieve (e.g., number ofmeasles doses administered, number of vitamin Acapsules distributed, etc.). However, programbaseddata do not reflect the extent of coverage ofthese programs (unless one estimates adenominator for the catchment area that convertsthese program statistics into a rate). Moreover,data from program participants are potentially8biased (do not reflect the situation of the generalpopulation), because of selectivity; that is, personswho opt to participate in the programs are oftendifferent from the population at large.In the conceptual framework described later in thischapter, output measures have been classified asprogram-based and outcome measures aspopulation-based. This classification is useful forevaluating national programs such as a nationalvitamin A program. However, it is less useful(especially the term “outcome”) for the evaluationof specific functional areas such as behavior changecommunication, training, and commodities andlogistics management. For example, one objectiveof training programs is usually improved qualityof service delivery. Although the collective effortsof training will contribute to outcomes at thenational level (e.g., improved family andcommunity child health practices), the most directand measurable effect of training is improvedservice quality. In this sense, the desired outcomefor a series of training events is quality of care in aspecific network of facilities. These results are notpopulation-based, but they represent theappropriate endpoint for monitoring andevaluating training programs. Thus, the “desiredoutcomes” for functional areas such asmanagement, training, logistics, and BCC areappropriately measured at the program level. Thisreport cannot resolve the debate about how toclassify the results of program efforts. It can onlyprovide guidance for indicator classification and arecommendation that programs specify theirdefinitions of the terms “input,” “process,”“output,” “outcome,” and “impact” in theirmonitoring and evaluation plans and frameworks.In contrast, governmental programs designed tohave national coverage are evaluated in terms oftheir effect on the general public. The term“population-based” can refer to a smallergeographic region (e.g., the catchment area for ademonstration project, such as a district), providedthe data are drawn from a representative sampleof the population.Chapter 1

B.4. Monitoring and Evaluation during Phases ofthe Program CycleChild health programs need to be evaluated atdifferent phases of the program cycle. Thesephases quite often overlap, and some programsmay skip certain phases altogether. However,thinking about the phases helps to provide anoverall picture of the program. It also helpsevaluators and program managers to conceptualizethe functions and decisions an evaluation is toserve, the kinds of questions it is to address, andthe data collection approaches that would beneeded to address different evaluation needs.Thinking about phases of the program cycle wouldalso help to structure the evaluation to facilitatethe use and impact of the findings. Figure 1.1(on page 10) provides a summary of the keymonitoring and evaluation issues for programs atdifferent stages (Herman et al., 1987).B.5. Value of a Conceptual FrameworkThe complexity and multiplicity of child healthinterventions make it difficult to capture all individualprogram components in one monitoring andevaluation framework. In some cases, differentprograms use different processes to arrive at the sameoutcome; in others, various programs use similarmeans to achieve different outcomes. A conceptualframework is useful for sorting out causal linkages—capturing the ways in which the processes/activitiesof the program affect the knowledge, attitudes, skills,behaviors of the target population. In this sense, aconceptual framework can help identify whatevaluation information might be most useful to theprimary intended users.The conceptual framework illustrated in Figure 1.2(on page 12) was designed to provide guidance onhow to define and select indicators for monitoringand evaluating child health programs. Theframework is adapted from a similar model developedunder the EVALUATION project (Bertrand andTsui, 1995). The framework is organized aroundthe standard input-process-output-outcome-impactschema and suggests a typical chain of programevents: inputs must be assembled to get the programOverviewunderway; activities are then undertaken withavailable resources; program participants engage inprogram activities; and as a result of what theyexperience, changes occur in knowledge, attitudes,and skills. Behavior and practice changes followknowledge and attitudinal changes, leading to theprogram outcomes, both intended and unintended.The conceptual framework was developed with anationally scaled program in mind. The frameworkcan be applied at a lower scale, such the regional ordistrict level, but the scale of the expected outcomesshould be adjusted accordingly. The shaded boxesrepresent the areas most commonly covered byroutine monitoring systems.Individual, household and community child healthoutcomes are influenced by many factors, animportant one being the broader context in whichprograms operate. This context includes the social,cultural and individual factors, including education,maternal health and nutrition, and genetic risk, manyof which are often outside the control of programs.Inputs include the financial and staff resources,equipment and supplies, treatment protocols, andessential drugs, and vaccines. Another criticalelement is the political and administrative system inwhich programs operate. The system influences howchild health is organized in a given country, theinfrastructure available for service delivery, the typeof service delivery strategies that are used (clinic basedor community-based, or and social marketing), andthe relative contribution of the public and privatesectors to that effort (Bertrand et al., 1995). Theframework also recognizes the contributions ofdonors to health service provision in many developingcountry settings.The inputs into child health programs are investedinto processes. Processes refer to the series of activitiesthat are carried out at the planning andimplementation phases of a program in order toachieve specific program objectives. Many activitiesare designed to achieve results outside the healthservices area. These activities seek to improvecapacity for the planning and management of childhealth services, improve health system support, and9

Figure 1.1. Program phases and priority monitoring and evaluation activities.Phase 1 - Program initiationEarly in the development of a new program or policy, sponsors and program managers consider thegoals they hope to accomplish through program activities and identify the needs or problems to beaddressed by the program. A needs assessment may be conducted at this stage to try and help structurea program. The questions addressed at this stage are the following: What needs attention? Whatshould the program try to accomplish? Data gathering at this stage can be used to make decisionsabout how to allocate money and effort in order to meet identified program needs and in order toprovide baseline data. The activities that follow from this stage are program planning or revision ofexisting programs.Phase 2 - Program planningThe program may be designed from scratch to meet the goals identified by a needs assessment or analready existing program may be revised or adapted to meet desired goals. During this phase, monitoringand evaluation activities may include the development of a monitoring and evaluation plan, controlledpilot testing and market testing to assess the effectiveness and feasibility of alternative methods ofservice delivery. The monitoring and evaluation plan should typically include the following components:assumptions regarding context, activities, and goals; anticipated relationships between activities, targets,and outcomes; well-specified measures and their operational definitions (indicators and metrics) andbaseline values; monitoring schedule, data sources, and M&E resource estimates; partnerships andcollaborations required to achieve results; and a plan for data dissemination and use.Phase 3 - Program implementationIn order to assess whether a program attains intended outcomes and meets participants' needs, it isessential to know what occurred in the program and that these activities can be reasonably connectedto outcomes. At this program stage, staff is trying to operationalize the program, adapt it as necessaryto a particular setting, solve problems that arise and get the program to a point where it is runningsmoothly. Monitoring and evaluation activities provide information that describes how the program isoperating and contributes to ways to improve it. Activities include continuous data collection such aswith service statistics, special studies to gather information not covered in a routine health informationsystem, qualitative studies to get in-depth insights into why the program may or may not beaccomplishing what it wants to accomplish.At the implementation stage, the questions addressed by monitoring and evaluation activities include:To what extent has the program been implemented as designed? How much does implementationvary from site to site? How can the program be improved? How can it become more efficient oreffective? As a result of these activities, revisions may be made to staffing, materials, activities, and theorganizational or administrative aspects of the program. Sometimes, the information may be used tomake decisions about the program based on whether or not the program's stakeholders think theactivities occurring will probably be effective in achieving other goals.Phase 4 - Program accountabilityAt this stage, the program has become established with a permanent budget and an organizationalniche and the purpose of evaluation is to assess the extent to which a program's highest priority goalsare and are not being achieved: To what extent has the program met its goals? Some of these goalsmight be satisfaction with the program, knowledge or skill gain, or behavioral. Activities at this stageinclude data reporting and dissemination and use of data for planning and management decisions.Decisions and actions likely to follow monitoring and evaluation activities at this stage are thoseconcerning whether to continue a program and in an expanded or reduced form.10Chapter 1

strengthen the policy environment. Strengthenedhealth system elements that are required to maintainboth facility and community-level activities includeimprovement in clinical supervision and logistics, thestrengthening or development of routine surveillanceand reporting systems, and improved capacity forplanning and management of child health services.At the policy level, program activities strive to createa supportive policy environment. This may includethe formulation of clear policies and guidelines, thelegal regulatory environment that affects productdevelopment, pricing and distribution, and theprovision of financial, material and human resourcesneeded by child health programs.At the program planning stage, processes may includecoalition building (for example, initiatives to bridgethe community and health-facility based services,dialogue between professional health serviceproviders and community-based health workers, andbetween public and private-sector providers),information-based needs assessment, formulationof a strategy for implementation, the assessmentof resource needs and availability, and theestablishment of monitoring and evaluationprocedures. At the program implementation stage,activities may include behavior changeinterventions, outreach, community mobilization,advocacy, training, logistics, supervision, and crosssectoralcollaboration.Many interventions are also designed tostrengthen political support and/or developeffective national policies in support of child healthprograms. Consequently, program implementationactivities at the process level may include theupdate and revision of existing clinical protocolsand training guidelines, policy development, andadvocacy. For example, where community IMCIhas not yet been adopted, programs may advocatefor the adoption of community IMCI as a nationalstrategy.The results achieved from the set of activities inwhich child health programs invest their humanand financial resources are called outputs. Outputsmay be defined for each of the functional areas ofOverviewa program. Functional outputs measure thenumber of activities conducted in each functionalarea such as capacity building, BCC, strategicplanning, and management. Functional areaoutputs may include, for example, the number ofhealth workers trained, the number of communitymeetings held, the number of IEC messagesdeveloped and disseminated, the number ofdistricts with micro plans, the existence of astrategic plan, and so forth.The outputs from the different functional areascontribute collectively to defining the serviceoutputs. Typically, programs strive to improve theadequacy of the service delivery system. Serviceoutputs can be classified and evaluated on threedimensions: access, quality (including referral andcounseling), and program image. In thisframework, quality is conceptualized both in termsof the technical performance of health workers,as well as the efficiency of service delivery,interpersonal relations, the continuity of services,physical infrastructure, and client satisfaction.Implicit in this conceptual framework are feedbackloops. The results obtained on output indicatorsmay require a reexamination of the activitiesundertaken by the program in different functionalareas and may require changes in program input.The boxes on the right side of the conceptualframework reflect the intended outcomes of childhealth programs. Child health programs striveoften to promote improvements in key behaviorsproven to be essential for child survival in theintermediate term and over the long term. Thekey outcome of child health programs is animprovement in the knowledge, attitudes, andchild health practices of caretakers, households andcommunities and coverage, which are critical forreducing overall child mortality. These outcomesare most relevant for the prevention componentof child health programs. At the household andcommunity levels many programs strive to increaseknowledge of preventive health behaviors, earlyrecognition of danger signs, and knowledge of sickchild management. Key child health practicesinclude the emphasis behaviors recommended by11

Figure 1.2. Conceptual framework for monitoring and evaluating child health programs.12

WHO/UNICEF for the prevention of illness athome, improved home management of childhoodillness and improved care-seeking for preventiveservices, such as vitamin A and immunization, andfor curative services in communities or healthfacilities. Improved child health practices andtimely care seeking behavior are measured at thepopulation level, and include compliance withtreatment recommendations (dosage and duration)and referral after receiving care from a healthworker; increased fluids and continued feedingduring illness; the provision of appropriatenutrition management (exclusive breastfeedingand complementary feeding); adequateconsumption of micronutrients; and ensuring thata child receives a full course of childhoodvaccination in the first year of life.The precise individual, household, and communitypractices that a program strives to change dependon its technical interventions. With respect tothe prevention and control of childhood diarrhea,for example, desired intermediate outcomes mayinclude proper hand washing at critical times;protection of drinking water from fecalcontamination; protection of food from fecalcontamination; use of ORS; and care seeking froma trained health provider when the child suffersfrom certain symptoms.By improving individual, household, andcommunity practices, and by making services moreaccessible and satisfactory to potential clients,programs may also strive to achieve increased andsustained demand for child health services andappropriate service use. The “appropriateness” ofservice use is emphasized for a number of reasons,an important one being the renewed emphasis onthe prevention and management of illness at homeand in the community. A case in point is theWHO Roll Back Malaria Strategy of “home asthe first hospital” which places increased emphasison caretaker recognition of malaria and treatmentseeking in both formal and non-formal health caresystems. Consequently, depending on the natureof the technical intervention, a general increase inthe utilization of facility-based services may or maynot be a measure of program success.Both improved individual, family, and communityhealth practices and the appropriate use of servicesare closely and directly linked with the long-termgoal of child health programs, which is to improveinfant and child health and nutrition and to reduceinfant and child mortality. While many programsare designed to reduce overall child mortality, itoften takes years to produce this result and it isnot always possible to make a causal link betweenthe child health program in question and mortalitydecline as it is also influenced by many nonprogramfactors (such as socio-economicconditions and the status of women).Consequently, program evaluations oftenconcentrate on outcomes that are more directlylinked to program effort and which are expectedto reflect change over a shorter period of time.As indicated in the upper right corner of theconceptual framework, institutionalization andprogram sustainability are also explicit goals ofmany donors and national/regional/district-levelchild health programs. Although successes in thisarea may not translate into gains in child healthin the short run, the extent to which programefforts have enhanced institutional capacity andprogram sustainability are legitimate foci ofevaluation efforts.B.6. IndicatorsCentral to program monitoring and evaluationefforts is the development of a set of indicatorsthat assess if and how well program activities havebeen carried out and whether program objectiveshave been achieved. An indicator may be definedsimply as a condition that can be empiricallymeasured. For example, the measles coverage rateprovides a measure of the extent to which measlesimmunization efforts have been successful inreaching the program’s target population.Many indicators have been developed for childhealth programs. However, to date most of theseOverview13

have not been compiled in a single document. Thegeneral objective of the guide is to bring togetherindicators that have already been tested and usedextensively in the child health field and compilethem in a single volume. By consolidating existingindicators and creating a framework formonitoring and evaluating child health programs,this guide aims to promote monitoring andevaluation by making indicators more readilyavailable to program managers and facilitating theuse of consistent terminology across programs,countries, and donor agencies.The indicators presented in the guide recapitulateand expand on indicators developed by CDC,UNICEF, USAID, and WHO for assessing healthprovider, household and community practices thataffect child health. Because of the effortsundertaken by these organizations to standardizechild health indicators, a broad consensus has beenreached on the best measures and data collectiontools for many of the indicators presented in theguide. For example, Chapter II draws heavily onthe core indicators established by WHO formonitoring and evaluating programs for theprevention of HIV in infants and young children.Chapter III presents indicators established byWHO and CDC for monitoring and evaluatingnewborn health at the global and national levels.The indicators presented in Chapter V weredeveloped over a long period of time by theIntegrated Disease Surveillance and ResponseTask Force, with considerable consensus-buildingby WHO-AFRO in collaboration with CDC,WHO-HQ and Ministries of Health in the Africaregion. Chapter VI of the guide presents a set ofpriority indicators for IMCI at the health facilitylevel developed by the Interagency WorkingGroup on IMCI, coordinated by WHO. Theseindicators reflect extensive field-based experienceand were selected to be valid, reliable,programmatically important, sensitive enough todemonstrate change, and measurable.The indicators presented in this guide are by nomeans exhaustive. Because of the complexity ofthe child health field and because substantial14resources can go into collecting data at the nationallevel, the number of indicators in anyprogrammatic area must remain limited. The setof indicators presented in this volume will notcomprehensively address all the specificmonitoring and evaluation needs of a nationalprogram in a given country or of individualprojects.B.7. How to Select IndicatorsIn general, the program’s objectives and phase ofimplementation, the evaluators’ role, and theinformation needs of the program’s managers andstakeholders will guide decisions about what tomeasure, observe or analyze. The four main stepsin selecting indicators are as follows:(1) State (or formulate) the objectives of theprogram;(2) Review the activities to be carried out inpursuit of the objectives;(3) Develop a simple framework to show how theprogram will work; that is, how the activitieswill lead to the desired objectives; and4) Select indicators that measure progress in eachof those boxes.Selecting indicators and setting targets is usuallydone during the process of program planning andreplanning, preferably in a participatory way withthe implementing agency and key stakeholders.While the level of attainment to be measured byan indicator is not usually part of the indicatoritself, it is a critical factor. The magnitude of thelevel to be measured affects the size of the sampleof the population needed to estimate that levelaccurately. It may also help select indicators thatmight assist in later interpretations of the result.The following questions can be helpful in selectingindicators:(1) Are program objectives measurable?(2) Are the data needed to measure the indicatorsavailable? If not, are they feasible to collect?(3) Are there alternative measures that need tobe considered?Chapter 1

(4) How often will the program report on thedifferent results? Will the data be availableby internal or external deadlines?(5) What financial support is available formonitoring and evaluation? Does theorganization have funds to conduct a survey?Or does the budget dictate the use of existingdata such as service statistics?(6) What are the requirements of the donoragency (if applicable)?Ideally, indictors should be:• Valid –They should measure the condition orevent they are intended to measure.• Reliable – They should produce the sameresults when used more than once to measurethe same condition or event, all things beingequal (for example, using the same methods/tools/instruments).• Specific – They should measure only thecondition or event they are intended tomeasure.• Sensitive – They should reflect changes in thestate of the condition or event underobservation.• Operational – It should be possible tomeasure or quantify them with developed andtested definitions and reference standards.• Affordable – The costs of measuring theindicators should also be reasonable.• Feasible – It should be possible to carry outthe proposed data collection.Validity is inherent in the actual content of theindicator and also depends on its potential forbeing measured. Reliability is inherent in themethodology used to measure the indicator andin the person using the methodology.In many areas of health, there is a tendency todevelop indices or composite/summary measuresthat encompass several areas of service provision.These summary indicators are useful in that theyOverviewlimit the number of statistics that need to bepresented at the highest policy level or to peoplewho are not specialists in the field and just need ageneral idea of whether things are getting betteror worse. The limitation of summary indicatorsis that changes are harder to interpret. A higherscore may mean an improvement across allcomponents measured by the index, or may be theresult of massive improvement in one area but anactual deterioration in another. Programmanagers, who need to know about theperformance of all components, will be interestedin disaggregated data that allow them to seeprogress in each area of service provision separately(UNAIDS, 2000). It is important to bear in mindthat aggregation too early in the process of datacollection or analysis may not meet the needs ofprogram or project managers.B.8. Data SourcesThere are a number of options for collectingessential data for monitoring and evaluating childhealth programs in developing countries. Theseare summarized briefly below.*B.8.1. Household and community level dataPopulation-Based Surveys: The most frequentlyconducted surveys are:USAID Demographic and Health Survey(DHS): Comprehensive large sample surveysthat include information on maternal andchild health, reproductive health, andmortality. A national sampling frame is usuallyused, although data can sometimes bedisaggregated to the level of smalleradministrative units such as districts. Thesesurveys provide useful background data foridentifying health priorities at the householdand community levels. These data can be alsoused for policy development and programplanning.* This section is adapted from the 2000 Technical ReferenceMaterials from the USAID/BHR/PVC Child SurvivalGrants Program, updated by Child Survival TechnicalServices Group, Macro International, Inc.15

UNICEF-Multiple Indicator Survey (MICS):Comprehensive large sample surveys thatinclude information on maternal and childhealth, reproductive health and mortality.These data can also be used for policydevelopment and program planning.30-Cluster Survey (WHO Control of DiarrhealDisease (CDD)/ARI/Breastfeeding Survey;WHO Immunization Coverage Survey/PVOKPC Survey): The 30-cluster methodology isoften used with reasonable precision by PVOsto obtain information in a project area. Surveyinstruments collect data on householdknowledge and practices for key maternal andchild health behaviors. Health behaviors areused as program outcome measures – thismethod is often used to collect baseline andfollow-up data.Rapid Catch: The Child SurvivalCollaborations and Resources Group(CORE) Monitoring and EvaluationWorking Group (MEWG) strongly suggeststhat organizations include all the RapidCore Assessment Tool on Child Health(CATCH) questions in their population-levelbaseline survey. The Rapid CATCH contains26 questions from the KPC2000+ modulesthat are considered important measures ofchild health. The Rapid CATCH has anaccompanying Tabulation Plan, which listspriority child health indicators and providesinstructions on calculating these indicators.The CORE MEWG suggests that USAIDsupportedchild survival projects report onthese priority indicators of householdbehaviors and care-seeking patterns that affectthe health and survival of children.Data from the Rapid CATCH can be used byimplementing PVOs and their local partnersto: (1) inform planning, monitoring, andevaluation activities; (2) provide a basis forcomparability between projects within a givencountry, as well as across countries; and (3)advocate at both the national and internationallevels for child health resources.Other methods for obtaining data include:Lot Quality Assurance Sampling (LQAS):LQAS uses small samples to determinewhether health behaviors are reaching predeterminedlevels. This approach is used insome settings to monitor the performance ofhealth services in small health areas. Themethod consists of the establishment of anacceptable and unacceptable performancethreshold, preferably with local input, andidentifying areas with poor performance inwhich remedial action needs to be taken. Thesurvey instruments are often the same as thoseused for larger sample surveys.Census-based Household Information Systems: Insome settings it may be possible to track allhouseholds in a community using regular visitsby trained workers. This system allows dataon vital events (births, deaths, pregnancies,episodes of illness) to be gathered, and alsoallows tracking of household knowledge andpractices – and the collection of healthindicator data. If regular visits are completeand sustained, then an accurate picture of thehealth status of a population can be obtained(since sampling is not required). Whenestablishing census-based systems, strategiesfor local use of data for planning purposes needto be developed, and strategies for sustaininghousehold visits over time need to beelaborated upon.Sample Vital Registration with Verbal Autopsy(SAVVY): This method is based on sentineldemographic surveillance and sample vitalregistration systems and uses a validated verbalautopsy tool to ascertain major causes of death.All births, deaths, and migrations at sentinelsites are enumerated through annual or semiannual“census-update” rounds. A SAVVYsystem can yield reliable national estimates ofall leading causes of death including HIV/AIDS, malaria, respiratory infection, diarrhealdisease, and maternal mortality. Data can beaggregated over multiple years to producerobust estimates for sub-national areas, agegroups, or poverty groupings.16Chapter 1

B.8.2. Health systems dataRoutine Health Information System Data: Thesedata are collected by facility-based staff andrecorded on standard reporting forms that are sentto higher levels in the system where they areaggregated. Data are most often service statisticssuch as the number of cases seen by category;number of deaths at the facility; number ofpregnancies and births; number of vaccinationsgiven, and the estimates of coverage using localpopulation data; and the number of outreach visitsconducted. These data may all be useful formonitoring or evaluating elements of programperformance. The advantage of this method is thatit uses routine systems and does not requireadditional resources. These data do not presentany information on health worker performance –a critical element of quality of care.Health Facility Surveys (HFS): These surveymethods include WHO integrated HFS, BASICSintegrated HFA, CORE adapted and integratedHFS, and USAID-SPA, and usually focus onoutpatient services at first-level and referralfacilities. Hospital-based care is not included.Facilities in the project area are sampled.Instruments need to be adapted, translated andpre-tested. Measures of health worker clinicalperformance for the management of key childhealth problems (ARI, diarrhea, malaria, measles,and nutrition) are collected (in the areas ofassessment, classification, treatment, andcounseling). Direct observations of practice arerequired, as well as exit interviews with caretakersof young children when they leave facilities.Health worker performance outcome measures areimportant measures of quality of care, and can beused to monitor improvements in clinical practice.Health facility system performance indicators arealso collected by health facility assessments, andinclude the availability of essential drugs, vaccines,supplies, equipment, and services; the availabilityof vehicles; the availability of health educationmaterials; and the infrastructure of the facility. Allare important measures of quality of care.Supervisory-based Data Collection: This method ismost often used to collect monitoring data, toprovide feedback to staff, and to engage these staffin solving problems locally. Structured checkliststhat include an observation of clinical practice areusually used. Instruments, methods, andindicators are similar to those used for healthfacility assessments. Supervisory methods mayallow a complete census of facilities in project areas,and therefore are an accurate measure ofperformance. This approach can also be used toobserve the practices of community-based healthworkers.Self-Assessment Methods and Peer-AssessmentMethods: Tools have been developed for self andpeer assessments of quality of care at the healthfacility level (COPE methods). The advantage ofthese approaches is that the cost and logisticaldifficulties of getting surveyors or supervisors tofacilities is eliminated. While these methods arevery useful in engaging local staff in the processof identifying and solving their own problems, theyare not considered valid or reliable enough for thepurpose of program evaluation. However, theycan play an important role in program monitoring.Program reviews: Program reviews are a systematicreview of key program elements using a structuredapproach. Evaluations are conducted using severalmethods, including reviewing routinedocumentation and reports; interviews with keyinformants; and structured checklists. Trainedstaff are needed to conduct reviews, and visits needto be made to each level. Program reviews canfocus at all levels of a system (national, provincial,district, sub-district, facility, community) or atsingle components (district-level). Programreviews might include reviews of routinesurveillance data; program plans and financialrecords; training records and lists of staff trained;the number of staff in facilities or communities;the availability of essential drugs and supplies; orthe regularity of supervision activities. Programreviews can also include activities at thecommunity level such as the availability of trainedOverview17

community health workers; the regularity ofoutreach visits; the regularity of health educationactivities; and the frequency with which healthplanning committees meet. Several elements ofsupport systems can be quantified and tracked overtime (such as the proportion of facilities withtrained staff; the proportion of districts with childhealth plans; and the proportion of the populationin a community with access to a trainedcommunity health worker).B.8.3. Qualitative dataAlthough qualitative indicators are not coveredin this guide, it is important to point out theirrole in monitoring and evaluation efforts.Qualitative data play an important role in impactevaluation by providing information useful tounderstand the processes behind observed resultsand assess changes in people’s perceptions of theirhealth status. At the population level, these dataare most useful for identifying health prioritiesand barriers to health practices; identifying localperceptions and beliefs about illness and theprevention and treatment of illness; andidentifying local terms for illness. These data arealso important for developing program strategies,adapting or improving the quality of M&Einstruments, and strengthening the design ofsurvey questionnaires.At the health systems level, qualitative data arecritical for identifying failures in various healthsystem components; the possible reasons for healthsystem failures; and possible strategies forimproving strategies in the local setting. At thehealth facility level, qualitative data are importantfor identifying perceptions of clients and healthstaff that may influence the delivery of care;identifying problems with care delivery, andidentifying solutions to these problems; definingquality care for the local population based oncultural and ethnic norms; investigating barriersto compliance with medications; identifyingbarriers to referral of sick children; and reviewingand improving counseling behaviors by healthworkers. Sometimes, information can often be18transformed into a quantitative rating scale againstwhich targets can be set.Qualitative methods most commonly used inmonitoring and evaluation can be categorized asfollows:In-depth Interviews: These tend to be openendedand to range from a total lack ofstructure and minimum control over aninformant’s responses to semi-structuredinterviews based on a written list of questionsand topics that need to be covered in aparticular order, and fully-structured interviewtechniques that may include pile sorting, frameelicitation, triad sorting, and tasks that requireinformants to rate or rank order a list of things(for more information about thesemethodologies, see Patton, 2002; Russell-Bernard, 1995). Open-ended questions andprobes are used to elicit information aboutrespondents’ experiences, perceptions,opinions, feelings, and knowledge.Focus groups are a particular type of in-depthinterview in which a group of people, usually6 to 12, and a moderator are recruited todiscuss a particular topic. Focus groups areless expensive to conduct than surveys andprovide insights on how people feel about aparticular product, issue, or behavior, and whythey feel that way.Observation: The methodology entailsfieldwork descriptions of activities, behaviors,actions, conversations, interpersonal reactions,organizational or community processes. Dataconsists of rich detailed descriptions thatinclude the context in which the observationswere made. A variety of data collectionmethods can be used. These includeobservations, conversations, interviews,checklists, and unobtrusive methods. Direct,reactive observations entail the investigatorengaging personally in all or part of theprogram under study or participating as aregular program member or client as aChapter 1

participant observer in order to gain greaterinsights than could be obtained from a surveyquestionnaire. Continuous monitoring andspot sampling can also be used. Nonreactiveand unobtrusive observation includes allmethodologies in which case informants donot know they are being studied (e.g., behaviortrace studies, archival research, contentanalysis, and two methods that pose seriousethical problems: disguised observations andnaturalistic field experiments). For furtherdetails on these methodologies, see Russell-Bernard (1995).Document Review: This methodology entailsstudying written materials and otherdocuments from organizational, clinical orprogram records, memoranda andcorrespondence; official publications andreports; personal diaries, letters, artistic works,photographs, and memorabilia; and writtenresponses to open-ended surveys. Data consistof excerpts, quotations, or entire passages fromthese documents that record and preserve thecontext.Rapid appraisals are becoming an increasinglycommon way of obtaining information on theneeds of the most vulnerable populations andinvolving these groups in decisions about their ownhealth improvements. This methodology is fairlyquick and cost-effective and addresses problemsof communities, rather than individuals. Methodsof data collection include ranking, mapping,diagramming, scoring, open interviews, andparticipant observation. There are different typesof rapid appraisals. Rapid rural appraisal involvesthe use of multidisciplinary teams to collect datafrom people in rural communities. Participatoryrural appraisals are a technique by which programmanagers, planners, and the community arepartners in information collection and analysis andin proposing solutions to the problems identified.Rapid epidemiological assessments involvesurveys, sampling, and risk assessments to evaluatehealth service functions. Rapid assessmentprocedures involve the use of anthropologicalmethods to assess community views of health,diseases, and health interventions. Rapidethnographic assessments involve the use ofanthropological methods to assess communitybeliefs and practices in relation to specific diseaseinterventions (Annett and Rifkin, 1995).B.9. Use of Indicators at Different LevelsMany commonly used indicators presented in theguide have grown out of international surveyprograms such as the Demographic and HealthSurveys and the Multiple Indicator ClusterSurveys, or out of protocols promoted byinternational donor organizations such as WHOand other United Nations agencies. Suchindicators permit comparisons between differentcountries and are useful at the international levelfor identifying regional trends or patterns in childmorbidity and mortality; highlighting persistentglobal and regional problems in child health;tracking trends in epidemics and response on aglobal scale; and allocating financial and technicalresources so as to have the greatest impact on childhealth. It is important, therefore, that indicatorsare defined and measured in the same way so thatthey can be compared directly from one countryto the next.The indicators can be used for similar purposes atthe national level. The use of comparable measurescan provide national programs with valuablemeasures of the same indicator in differentpopulations, enabling triangulation of findings andallowing regional or local differences to be notedand addressed. This can help to direct resourcesto regions or subpopulations with greater healthneeds and identify areas for intensification orreduction of effort (UNAIDS, 2000a). In decidingon a national set of indicators, it is important thatcountries realize that they are not limited to theset of indicators described in this volume, norshould they necessarily collect all of them. Thechoice of indicators should be driven by theobjectives, goals, and activities of the nationalprogram, taking into consideration the time andmoney it costs to collect and analyze data for eachOverview19

indicator. Many of the indicators presented in thisguide are core indicators that have beenrecommended for collection in all countries byinternational health organizations such as CDC,UNICEF, and WHO. Supplemental indicatorsfor certain programmatic areas, such as IMCI andnewborn health, are only presented in a fewinstances. Where they fit the needs of a country,national programs are encouraged to use theindicator definitions presented here to ensurestandardization of information over time. Manyof the indicators listed in the guide are also suitablefor district monitoring and evaluation purposes.Ideally, indicators should be measured with datathat are already available. It should be possible touse data from routine health reporting systems toobtain data for calculating some output andcoverage indicators. Frequently, data for manyoutcome and impact indicators will need to becollected through health facility or populationbasedsurveys. The cost and benefits of variousdata collection options must be borne in mindwhen indicators are chosen to measure change inareas of program effort.At the project level, the choice of indicatorsdepends on what the project wants to do. Whilemany of the indicators in this volume are relevantfor use in the population at large or in healthservice settings, they do not cover the full rangeof monitoring and evaluation needs for specificprojects. Some projects may find some of theindicators contained in the guide at odds with theirown operational definitions. In such cases,collecting and reporting on data in the wayspecified in the guide may or may not meet theproject’s information needs. If a measurementmethod comparable to one proposed in this guideis being used or if the project’s monitoring andevaluation activities can be modified slightlywithout compromising the evaluation of theproject, then those indicators which are relevantto the project should be collected and reported.The use of uniform definitions can provide thenational program with comparable measures fordifferent populations and facilitate district andregional-level comparisons within and acrosscountries.20Chapter 1

RefererencesencesAnnett HB, & Rifkin SB. (1995). Guidelines for Rapid Participatory Appraisal to Assess CommunityHealth Needs: A Focus on Health Improvements for Low-Income Urban and Rural Areas. WHO/SHS/DHS/95.8. Geneva, Switzerland: World Health Organization.Bertrand JT, & Escudero G. (2002). Compendium of Indicators for Evaluating Reproductive HealthPrograms. Chapel Hill, NC: MEASURE Evaluation Project.Bertrand JT, Magnani RJ, & Ruttenberg N. (1994). Handbook of Indicators for Family Planning ProgramEvaluation. Chapel Hill, NC: The University of North Carolina at Chapel Hill, Carolina PopulationCenter, Evaluation Project.Bertrand JT, & Tsui A. (1995). Indicators for Reproductive Health Program Evaluation. Chapel Hill,NC: The University of North Carolina at Chapel Hill, Carolina Population Center, Evaluation Project.Bryce J, Victora CG, Habicht JP, Vaughn JP, & Black RE. (2004). The Multi-Country Evaluation ofthe Integrated Management of Childhood Illness Strategy: Lessons for the Evaluation of Public HealthInterventions. American Journal of Public Health, 94(3):406-415.Habicht JP, Victora CG, & Vaughan JP. (1999). Evaluation Designs for Adequacy, Plausibility andProbability of Public Health Programme Performance and Impact. International Journal of Epidemiology,28(1):10-8.Herman JL, Fitz-Gibbon LL, & Morris CT. (1987). Evaluator's Handbook. Thousand Oaks, CA:Sage Publications, Inc.Patton MQ. (2002). Qualitative Research and Evaluation Methods - Third Edition. Thousand Oaks,CA: Sage Publications, Inc.Russell BH. (1995). Research Methods in Anthropology: Qualitative and Quantitative Approaches - SecondEdition. New York: Altamira Press.Schalock RL. (1996). Outcome-based Evaluation. New York and London: Plenum Press.UNAIDS. (2000). Local Monitoring and Evaluation of the Integrated Prevention of Mother to Child HIVTransmission in Low-income Countries. Geneva, Switzerland: UNAIDS.UNAIDS. (2002). Monitoring the Declaration of Commitment on HIV/AIDS: Guideline on Constructionof Core Indicators. Geneva, Switzerland: UNAIDS.UNGASS. (2001). Declaration of Commitment on HIV/AIDS: Global Crisis - Global Action. UNGASS;United Nations General Assembly Twenty-Sixth Special Session.United States Agency for International Development, Bureau for Humanitarian Response, and Officeof Private and Voluntary Cooperation. (2000). PVO Child Survival Grants Program: Technical ReferenceMaterials. Washington, DC: USAIDOverview21

Wasantwisut E. (2002). Recommendations for Monitoring and Evaluating Vitamin A Programs: OutcomeIndicators. Journal of Nutrition, 2940S-2S.WHO. (1999). IMCI Indicators, Monitoring and Evaluation. Geneva, Switzerland: World HealthOrganization.WHO. (2000). Roll Back Malaria: Framework for Monitoring Progress and Evaluating Outcomes and Impact.Geneva, Switzerland: World Health Organization.22Chapter 1

Annex 1.1. A complete listing of goals and targets for the Millennium Development GoalsChapter 1. Overview23

2PREVEVENTIONOFMOTHEOTHER-TO-C-CHILD-CTRANSMISSION(PMTCTCT) OF HIVIndicators:• Existence of guidelines for the prevention of HIV infection in infantsand young children• Number and percentage of health care workers newly trained orretrained in the minimum package during the preceding 12 months• Prevention and care service points• Women completing the testing and counseling process• Percentage of HIV-positive pregnant women receiving a completecourse of ARV prophylaxis to reduce MTCT in accordance with anationally approved treatment protocol (or WHO/UNAIDS standards)in the preceding 12 months• Percentage of HIV-infected infants born to HIV-infected mothers·

HAPTER 2. P-CCHAPTER2. PREVENTIONOF MOTHERTHER-TO-C-CHILDTRANSMISSION(PMTCT)OF HIVBy the end of 2004, UNAIDS estimated that2.2 million children under 15 years of agewere living with HIV. In the year 2004 alone, itwas also estimated that 640,000 children werenewly infected with HIV and 510,000 died ofHIV/AIDS (UNAIDS, 2004). Almost all HIVinfectedchildren acquired HIV via mother-tochildtransmission (MTCT). Such transmissioncan occur during pregnancy, labor or delivery, andafter birth through breastfeeding. In the absenceof anti-retroviral intervention, it is estimated thatMTCT rates range from 5-10% during pregnancy,10-20% during labor and delivery, and 15-30% inthe absence of breastfeeding, to 25-35% withbreastfeeding through 6 months and to 30-45% ifthere is breastfeeding through 18 to 24 months(De Cock et al., 2000).At the United Nations General Assembly SpecialSession (UNGASS) on HIV/AIDS in June 2001,governments from 189 countries adopted theDeclaration of Commitment on HIV/AIDS. TheDeclaration established specific goals on a numberof quantified and time-bound targets, includingreductions in HIV infection among infants. TheUNGASS goal for prevention of HIV infectionsin infants and young children is presented below:“By 2005, reduce the proportion of infantsinfected with HIV by 20 percent, and by 50percent by 2001, by ensuring that 80% ofpregnant women accessing antenatal care haveinformation, counseling and other HIVprevention services available to them,increasing the availability of and by providingaccess for HIV-infected women and babies toeffective treatment to reduce mother-to-childtransmission of HIV, as well as througheffective interventions for HIV-infectedwomen, including voluntary and confidentialcounseling and testing, access to treatment,especially antiretroviral therapy (ART) and,where appropriate, breast-milk substitutes andprovision of a continuum of care” (UNGASSDeclaration of Commitment on HIV/AIDS, 2001,paragraph 54).The WHO framework for action to prevent HIVinfection in infants and young children includesthe following four strategies:• Primary prevention of HIV infection in allwomen (including the promotion ofabstinence, monogamy, and condom use).• Prevention of unintended pregnancy amongHIV-infected women (including theprovision of testing and counseling in familyplanning (FP) and other reproductive healthservices).• Prevention of HIV transmission from HIVinfectedwomen to their infants and youngchildren (including the use of antiretroviral(ARV) drugs, safer delivery practices, andcounseling and support for infant feeding).• Provision of care and support to HIVinfectedwomen and their infants andfamilies (including the prevention andtreatment of opportunistic infections, theuse of antiretroviral drug therapy (ART),psychosocial and nutritional support, andreproductive health care, including saferdelivery, FP, and counseling and support forinfant feeding).Many countries are expanding their programs forprevention of HIV infection in infants and youngchildren to respond to the growing HIV/AIDSpandemic and increased commitment and support.However, many programs remain costly relativeto per capita spending on health in many countries,so careful monitoring and evaluation of the successof interventions to reduce transmission of HIVfrom mothers to children is important. UNAIDSChapter 2. Prevention of Mother-to-Child Transmission (PMTCT) of HIV25

and its partners have developed a set of coreindicators that permit monitoring of keyinternational and national actions, nationalprogram outcomes, and impact. These arepresented in National AIDS Programmes: A Guideto Monitoring and Evaluation, which represents thejoint efforts of UNAIDS and multiple partnerorganizations (UNAIDS, 2000a). However,though this document was produced in recentyears, some important areas are not included,reflecting the rapid developments in HIV/AIDSprevention and care in the last few years. One ofthese key gaps is the insufficient attention paid tothe monitoring and evaluation of programs forreducing MTCT of HIV/AIDS.Ongoing monitoring and evaluation effortsinclude the development of guidelines for LocalMonitoring and Evaluation of the IntegratedPrevention of Mother to Child HIV Transmissionin Low-Income Countries, a joint collaborativeeffort by UNICEF, UNAIDS and WHO(UNAIDS, 2000b). These guidelines include: (1)locally monitoring the progress in implication,identifying problems, troubleshooting andadapting implementation strategies; (2) evaluatingthe effectiveness, impact, cost-effectiveness andfinancial sustainability of the intervention in pilotprojects; and (3) conducting applied research toaddress unresolved issues, test strategies foroptimizing the effectiveness, impact, costeffectivenessand financial sustainability, andminimizing the risks of the intervention programs.The guidelines also give advice on how to chooseindicators for monitoring, evaluation, andoperations research; establish methodologies toanalyze and use the information; and establishstandards for information systems. Localmanagers and planners are the primary audiencefor these guidelines.In 2004, WHO published the National Guide toMonitoring and Evaluating Programmes for thePrevention of HIV in Infants and Young Children(WHO, 2004). This national guide complementsexisting M&E guides, including Monitoring theDeclaration of Commitment on HIV/AIDS:Guidelines on Construction of Core Indicators(UNAIDS, 2002), which included two indicatorson programs for the prevention of HIV infectionin infants and young children. The national guidealso supplements National AIDS Programmes: AGuide to Monitoring and Evaluation, and like thatmanual, is intended for use by national MTCT,reproductive health, and HIV/AIDS programmanagers. The purpose of the national guide isto determine the level of success of programs forthe prevention of HIV infection in infants andyoung children, to identify areas where furthersupport is required, and to inform adaptation andscaling up strategies.Methodological Challenges of Monitoringand Evaluating PMTCT ProgramsPrevention of HIV in infants cuts across a number ofother programs, sometimes well integrated, sometimes not.In spite of recent efforts on M&E for HIV/AIDS,many challenges remain. Prevention of motherto-childtransmission can be considered as acascade of program components as follows:primary HIV prevention activities and VCTservices during antenatal care; improvement inbasic obstetrical care including offer of ARVs toHIV-positive pregnant women and adequatedelivery practices; counseling for infant feedingduring antenatal care (ANC); postpartum careincluding support to infant feeding, growthmonitoring, family planning services, andscreening of HIV infection in children; and longtermsupport to HIV infected mothers and theirfamilies. The wide range of interventions is achallenge to monitor and evaluate. Some of theseservices are well integrated into MCH servicesbut others are not. While some services are doneat the level of MCH services, others are done atthe community level (i.e. communicationprograms) or at the level of other health, socialservices, or NGOs (i.e. long-term care). Effectivemonitoring and evaluation of PMTCT serviceswould need to incorporate standard indicators thathave been used in the context of other programs.Some of these standard indicators includeantenatal care coverage, births attended by skilled26 Chapter 2

health personnel, availability of basic essentialobstetric care, family planning method mix,condom availability, and iron/folic acidsupplementation during pregnancy. Clearly,indicators that are selected for monitoring andevaluation of PMTCT interventions woulddepend on program goals.PMTCT lacks a set of standard, easily measurableindicators at the program output level.Standard monitoring and evaluation approachesuse indicators at three levels: the program level,outcome level, and impact level. Few standardindicators have been established for the M&E ofPMTCT programs at the program output andoutcome levels. Recent efforts to establish coreindicators for monitoring the United StatesPresident's Emergency Plan for AIDS Relief haveincluded functional output indicators for PMTCT.However, this is a work in progress and many ofthe proposed output indicators have only recentlybeen field-tested.Monitoring and evaluating replacement feeding is acomplex programmatic issue.Promotion of breastfeeding as the best possiblenutrition for infants has been the cornerstone ofchild health and survival strategies for the pasttwo decades, and has played a major part inlowering infant mortality in many parts of theworld. Breastmilk substitutes if used incorrectlyor over-diluted can cause infection, malnutrition,and death. Therefore, assessment of theeffectiveness of the infant feeding component ofPMTCT programs needs to be made in the lightof other indicators such as the availability of infantformula; access to clean water and fuel to boil it;and the ability of mothers/caregivers to make upreplacement feeds correctly. A promotion ofreplacement feeding among HIV-infected womenmay also lead to a reduction in the fertilityinhibitingeffects of breastfeeding, making theavailability of FP services a necessary part ofpostpartum care.Impact indicators are especially difficult to obtain.Regarding impact indicators, it is difficult inpractical terms to obtain information from a trulyrepresentative sample in a given country.Consequently, the current approach is to examinewomen attending antenatal clinics with theassumption that they represent a wide crosssectionof the population. However, in manycountries, large numbers of pregnant women withHIV may not have access to ANC services or maychoose not to utilize them.One measure of impact, infant infection status, isespecially difficult to obtain. HIV-testing at birthis of limited use for establishing the infection statusof infants because nearly half of all verticaltransmission in developing countries takes placein the postnatal period, during breastfeeding.Follow-up would be nearly impossible for routinesurveillance systems. In many countries,particularly those with high pre-AIDS mortalityin the under-fives and poor vital registrationsystems, infant and child mortality indicators arenot specific enough to register changes in rates ofHIV-associated mortality in infants.Selection of IndicatorsThis chapter draws on the core indicatorsestablished by WHO (2004) for monitoring andevaluating programs for the prevention of HIV ininfants and young children. These indicators arerecommended for M&E in all countries, regardlessof the type of epidemic. Some are newlydeveloped, whereas others have been used forseveral years. We have modified the format usedin the original document to be consistent with thatused in this guide. Otherwise, the content of thesecore indicators remains true to the original.The indicators included in this chapter are thefollowing:Output indicators• Existence of guidelines for the prevention ofHIV infection in infants and young childrenPrevention of Mother-to-Child Transmission (PMTCT) of HIV27

• Number and percentage of health careworkers newly trained or retrained in theminimum package during the preceding 12monthsOutcome indicators• Prevention and care service points• Women completing the testing andcounseling processImpact indicators• Percentage of HIV-positive pregnantwomen receiving a complete course of ARVprophylaxis to reduce MTCT in accordancewith a nationally approved treatmentprotocol (or WHO/UNAIDS standards)in the preceding 12 months• Percentage of HIV-infected infants born toHIV-infected mothersAs was pointed out by WHO (2004), the twoimpact indicators were established by UNGASSfor monitoring progress towards the achievementof specific targets established in the Declarationof Commitment on HIV/AIDS (UNGASS,2002b).The set of PMTCT indicators presented here doesnot comprehensively address all the specificmonitoring and evaluation needs of PMTCTprograms. Indicators of service provision forPMTCT should ideally include provision of VCTservices for pregnant women, the availability andaffordability of ARVs, provision of counseling andsupport for infant feeding, and the availability andaffordability of alternatives to breastmilk. In addition,indicators should measure crosscutting themesessential for program development and scaling upsuch as policy development and development ofhealth system capacity (WHO, 2004).At the project level, good monitoring andevaluation requires indicators that are directly tiedto project activities, goals, and objectives.Additional indicators are proposed in other M&Eguides for countries in which they are relevant tothe national epidemic or national response. TheUNAIDS guidelines for Local Monitoring andEvaluation of the Integrated Prevention of Motherto-ChildHIV Transmission in Low-IncomeCountries (UNAIDS, 2000b), for example, may beof greater relevance for monitoring and evaluatingspecific projects aimed at demonstrating thefeasibility and effectiveness of integratingPMTCT activities in routine MCH serviceswithin developing countries. The National Guideto Monitoring and Evaluating Programmes for thePrevention of HIV in Infants and Young Children(WHO, 2004) also presents additional indicatorsthat can be used at the national level dependingon programmatic needs and available resources.TerminologyThe term mother-to-child transmission of HIV(MTCT) is often used to refer to the transmissionof HIV to infants. As in the National Guide toMonitoring and Evaluating Programmes for thePrevention of HIV in Infants and Young Children(WHO, 2004), we use MTCT to refer to thebiological process of vertical transmission. Theterm prevention of mother-to-child transmission(PMTCT) is used to refer to the broad range ofrecommended strategies for the prevention of HIVinfection in women, infants, and young children,including, but not limited to, the provision ofARVs to HIV-infected women to prevent MTCT.28 Chapter 2

EXISXISTENTENCEOF GUIDELINESFORTHE PREVENTIONOFHIV INFECTIONIN INFNFANTANTS AND YOUNOUNG CHILDRENHIV IWHO PMTCT Core e IndicatorDefinitionExistence of national guidelines (either approvedor in draft form) for the prevention of HIVinfection in infants and young children and thecare of infants and young children in accordancewith international or commonly agreed standards.Guidelines should be available for all fourcomponents of the comprehensive strategy forpreventing HIV infection in infants and youngchildren; these components are as follows:(1) Intensified prevention efforts aimed at youngwomen (this may or may not be specificallyincluded in these guidelines and may beaddressed elsewhere);(2) Prevention of unintended pregnancies amongHIV infected women;(3) Specific interventions to prevent HIVtransmission from infected mothers to theirchildren, including ARVs; safe deliverypractices; and counseling and support forinfant-feeding; and(4) Referral or care for HIV-infected mothers andtheir children.In some countries, each of these issues may beaddressed as part of comprehensive national HIV/AIDS guidelines. In others, individual guidelinesmay be available. Information on HIV and infantfeeding may be incorporated in national guidelineson the feeding of infants and young children.Measurement ToolsA survey among key informants at the nationallevel or in health care facilities is used to determinewhether there are guidelines for each interventionprong. The key informants at the national levelare persons responsible for HIV/AIDS, maternaland child health (MCH) or infant feeding andnutrition; at the health facility level the keyinformants include practitioners and clinicdirectors. The actual guidelines, with evidence ofapproval or submission for approval, may also bereviewed.What It MeasuresNational guidelines are commonly based onexisting international standards, or on standardsabout which there is general agreement but whichhave not yet been formally presented asinternational guidance. Without guidelines,services of unknown quality and impact can beimplemented on an ad hoc basis, making it difficultto monitor and evaluate efforts.How to Measure ItMeasurement of this indicator requires evidenceof guidelines. These include guidelines to preventinitial infection among HIV-negative women;prevent unintended pregnancies among HIVpositivewomen; provide ARV prophylaxis, use safedelivery practices, and provide infant feedingcounseling and support among HIV-infectedpregnant and lactating women; and refer to orprovide care and support to HIV-infected womenand their children. When asking if such guidelinesexist, the following additional questions may beasked if time and resources allow:• How were these guidelines formulated?(Explore the process: ask by whom and onwhat basis they were formulated)• Are these guidelines nationally accepted (evenif only draft versions are available)?• To what extent are they implemented?(Explore the extent of implementation and thePrevention of Mother-to-Child Transmission (PMTCT) of HIV29

arriers and opportunities that were or arebeing encountered in implementation)• How often and by whom are they updated?The indicator should be measured and the abovequestions answered for each intervention asoutlined in the indicator’s definition. Thisindicator should be measured every year untilguidelines are found to exist.Strengths and LimitationsThis indicator is not concerned with the qualityof guidelines or that of their implementation.Furthermore, because it does not capture newdevelopments in the field, the guidelines have tobe reassessed periodically in order to guaranteethat they remain consistent with changingstandards.30Chapter 2

NUMBERAND PERERCENTCENTAGEOF HEALEALTHCAREWORKERSNEWLEWLY TRAINEDOR RETRAINEDIN THE MINIMUMPACKAKAGEDURING THE PRECEDIN12 MONTHSRECEDING 12 MWHO PMTCT Core e IndicatorDefinitionThe number and percentage of health care workersnewly trained or re-trained in the minimumpackage during the preceding 12 months.Numerator: Number of health care workersnewly trained or re-trained in the minimumpackage during the preceding 12 months.Denominator: Total number of health careworkers working in facilities that haveimplemented the minimum package, withwomen that could benefit from it, forpreventing HIV infection in women andinfants.“Re-trained health care workers” are those that haveundergone in-service training, i.e., they are alreadyin the work force and have been practicing forseveral years. Training includes both in-service andpre-service training.The “minimum package” varies between differenttypes of health care facilities. Several kinds offacilities that may provide services for preventionof HIV infection in infants and young childrenare outlined below, together with services thatshould be available. Providers working in thesesettings should be trained in each of thecomponents mentioned.ANC /MCH clinics: counseling on risk reduction,counseling on infant feeding, and referral orprovision of the following: HIV counseling andtesting, ARV prophylaxis, FP (includingcounseling on dual protection), attended deliveryin birth facilities where safe obstetric practices areobserved, and long-term care.FP clinics: counseling on dual protection; referralor provision of HIV counseling and testing, andlong-term care; and referral to ANC/MCHservices if appropriate.Maternity hospitals: observance of safe obstetricpractices, and referral to or provision of thefollowing: HIV counseling and testing, ARVprophylaxis, counseling on infant feeding, MCHservices, FP, and long-term care.Primary health care (PHC) facilities: referral to HIVcounseling and testing, ANC/MCH, and FPservices.Measurement ToolsTraining records; health facility surveyWhat It MeasuresThis indicator quantifies the human resources thatare trained in preventing HIV infection in womenand children and are available to provide therequired services. For the purpose of planning, itis important to assess the resources available toaddress health needs. Before the implementationor expansion of services, it is vital to know notonly what facilities and equipment are available,but also what training and human resources exist.Only with this information can health systemsprovide services that meet the needs of and areacceptable to the populations concerned.How to Measure ItThe numerator can be calculated on the basis of areview of training records in each facility that hasimplemented services or serves women who couldbenefit from the minimum package for preventingHIV in infants and young children. If however,such records do not exist, a survey of facilities canbe carried out. A random sample of health carePrevention of Mother-to-Child Transmission (PMTCT) of HIV31

providers in these facilities should be asked whattraining they may have received in the preventionof HIV infections among infants and youngchildren. (In some countries, a national, provincialor district training coordinator keeps records ofthe training given to individual health workers.Such data can be used instead of a facility survey.)Interviewers should investigate the compositionof the training, which varies with the type of site.The minimum package for each type of facility isoutlined in the definition of the indicator.The denominator, i.e., the number of staff able toprovide preventive services for HIV infectionamong infants and young children, is calculatedon the basis of the number of health care providersworking at sites where women could potentiallyreceive the services included in the minimumpackage. These data can be obtained from ministryand health facility records.trained. The indicator should be interpreted inrelation to the size and nature of the epidemic inparticular countries.Difficulties may occur in determining thedenominator, as some countries may have limitedinformation on the pool of human resourcesavailable in various facilities. Frequent transfersof personnel between facilities, or high rates ofattrition, may complicate the interpretation of thisindicator. It should be noted that the assumptionis made that only formal health workers arecounted, i.e. those remunerated either financiallyor in kind. In many settings, however, informalhealth workers make a significant contribution.The numerator should be collected every year. Thedenominator, if based on facility surveys, is moreexpensive, but is necessary for the calculation ofthe percentage and should be obtained every twoyears. After the initial collection of data, it maybe of interest to disaggregate data for those healthcare workers who have been newly trained orretrained during the preceding 12 months, as wellas to maintain a record of how many health careworkers have been trained since the first time thisindicator was measured.Strengths and LimitationsThis indicator is useful in that it tracks the numberof health workers trained to provide services forthe prevention of HIV infection in infants andyoung children over time. It attempts to documentincreasing capacity to deliver preventioninterventions. However, no conclusion should bedrawn regarding quality, because this is affectedby the practices employed rather than by theexistence of trained personnel. It should not beexpected that all health workers in countries willhave been trained, nor even that a high percentageof those who could be trained will have been32Chapter 2

PREVENTIONAND CARESERERVICEPOINTOINTSWHO PMTCT Core e IndicatorDefinitionThe percentage of public, missionary, andworkplace venues (FP and PHC clinics, ANC/MCH, and maternity hospitals) offering theminimum package of services for the preventionof HIV infection in infants and young children inthe preceding 12 months.Numerator: Number of public, missionary, andworkplace venues (FP and PHC clinics,ANC/MCH, and maternity hospitals)offering the minimum package of services forthe prevention of HIV infection in infants andyoung children in the preceding 12 months.Denominator: Total number of public,missionary, and workplace venues FP andPHC clinics, ANC/MCH, and maternityhospitals).The “minimum package” is defined, as with theprevious indicator, by the type of clinical setting.An outline is given below of facilities that mayprovide services for the prevention of HIVinfection in infants and young children, togetherwith indications of the services that should beavailable.ANC /MCH clinics: counseling on risk reduction,counseling on infant feeding, and referral orprovision of the following: HIV counseling andtesting, ARV prophylaxis, FP (includingcounseling on dual protection), attended deliveryin birth facilities where safe obstetric practices areobserved, and long term care.FP clinics: counseling on dual protection; referralor provision of HIV counseling and testing, andlong-term care; and referral to ANC/MCHservices if appropriate.Maternity hospitals: observance of safe obstetricpractices, and referral to or provision of thefollowing: HIV counseling and testing, ARVprophylaxis, counseling on infant feeding, MCHservices, FP, and long-term care.Primary health care (PHC) facilities: referral to HIVcounseling and testing, ANC/MCH, and FPservices.Measurement ToolsSurvey of all public, missionary, and workplacehealth facilities offering FP and PH, ANC/MCH,and maternity servicesFor overall coverage, the following instruments canbe adapted:• WHO draft protocol for the evaluation ofHIV/AIDS care and support (WHO, 2000)• UNAIDS protocol for evaluation of care andsupport (UNAIDS, 1996)These instruments involve the performance ofsurveys of health facilities.For HIV testing and counseling, the following toolcan be adapted:• UNAIDS tool for evaluating HIV voluntarycounseling and testing (UNAIDS, 2000c)What It MeasuresIn order to be effective, the prevention of HIV ininfants must be applied as broadly as possible atall relevant treatment sites. It is generallyacknowledged that a large proportion of preventiveservices occur in ANC settings. As servicesbecome integrated, however, it will be importantPrevention of Mother-to-Child Transmission (PMTCT) of HIV33

to consider other outlets as well. This indicatormeasures the coverage of the services at each ofthe outlets where prevention or care opportunitiesarise. These opportunities comprise either referralsto other services or the provision of on-siteservices.How to Measure ItThe information required for this indicator canbe obtained by various methods and depends onresource availability and the amount of detailsought. The first option requires that aquestionnaire be sent to all public and private FPand PHC clinics, ANC/MCH, and maternityservices. It should be facility-specific, outliningthe specific services on offer. A column should beincluded to show whether services are providedon-site or if referrals are made. If the number ofpossible sites to be surveyed is too great for all tobe covered, a stratified random sample, eachstratum being a different type of service delivery,and the questionnaire can be sent to the selectedsites.more common outlets for prevention and careamong women and infants.The indicator should be measured every two tothree years.Strengths and LimitationsThis indicator provides critical information on thenational availability of prevention and care effortsamong women and infants. While it is useful toprogram planners seeking to determine whereservices are needed, or where facilities areproviding the full spectrum of services to preventHIV infection in women and infants, it cannotmeasure the quality of the services being providedin each facility. Moreover, not all countries shouldbe expected to have all, or a high percentage of allpossible, health care service points offering servicesto prevent HIV in infants and young children.Rather, this indicator needs to be interpreted inlight of the size and nature of the epidemic acountry is facing.In measuring this indicator, special attention mustbe taken that the type of service providing theinformation is noted (i.e., ANC, family planningcenter, etc.). Only this way will one be able todetermine the more common outlets forprevention and care among women and infants.The availability of PMTCT services may also beanalyzed by geographic area, or by sector (publicor private).Scoring should not be done in an “all or nothing”fashion. The numerator should reflect thoseelements of the package that are present (or forwhich there are in-house referral mechanisms).The denominator is all public, missionary andworkplace venues (FP and PHC clinics, ANC/MCH, and maternity hospitals).Irrespective of the method adopted for measuringthe indicator, it is essential to note the type ofservice providing the information (e.g., ANC, FPcenter). This makes it possible to determine the34Chapter 2

WOMENCOMPLETINOMPLETING THE TESESTINTING AND COUNSELINOUNSELINGPROCESSDefinitionThe percentage of pregnant women making atleast one ANC visit who have received an HIVtest result and post-test counseling. **An additional important program-level counseling and testingindicator must be considered when a PMTCT program is beingmanaged. The indicator measures the points in the provisionof counseling and testing for pregnant women at which womendrop out. This information can be used to investigate furtherwhy women drop out at specific points and, ultimately, to reducethe percentage of drop-outs. Such information is thereforeimportant for program planning. The indicator includes thefollowing three components:(a)(b)(c)WHO PMTCT Core e IndicatorNumerator: Number of pregnant women whohave received an HIV test result and post-testcounseling in the preceding 12 months.Denominator: The estimated number ofpregnant women giving birth in the preceding12 months who have made at least one ANCvisit.Measurement ToolsReview of program recordsThe number of pregnant women who have made at leastone ANC visit and have been counseled at a PMTC site,divided by the total number of pregnant women.The number of pregnant women who have acceptedtesting for HIV, divided by the total number of pregnantwomen who have made at least one ANC visit and havebeen counseled at a PMTCT site.The number of women receiving post-test counseling andHIV results, divided by the total number of pregnantwomen who have made at least one ANC visit and havebeen counseled at a PMTCT site.What It MeasuresFor PMTCT to be effective, it is necessary to knowa woman’s sero-status in order to tailor preventionand care to her needs. A successful PMTCTprogram will reach as many pregnant women aspossible to ensure knowledge of sero-status. Thisindicator provides a broad measure of programcoverage in the country concerned. However,issues of poor access to services and poor uptakeresult in only a small percentage of womenknowing their status. It is therefore important torefer to the program-level indicator described inthe footnote at the bottom of this page.How to Measure ItThis indicator requires that program records bereviewed in order to count how many womencomplete each stage of the testing and counselingprocess, i.e. have received their test results andpost-test counseling. The number of women whohave made at least one ANC visit is estimated bymultiplying the number of births in the preceding12 months, as given in a census or the best availablesources, by the rate of ANC attendance (DHStypesample survey). In some cases, the numeratormay be obtained by examining national records.If this is not possible, the required data are likelyto be available at the district level, where they canbe collected directly from facilities providing theservices in question.In some cases, the denominator may be obtainableby examining national ANC registries. This isthe preferable denominator and should be used ifpossible. If this number is not available or reliable,the estimate of the number of pregnant womendescribed above can be substituted but thisapproach involves an increased possibility ofmisinterpretation.This indicator should be measured annually.Prevention of Mother-to-Child Transmission (PMTCT) of HIV35

Strengths and LimitationsAs stated in National AIDS Programmes: A Guideto Monitoring and Evaluation (UNAIDS, 2000),this indicator provides a broad measure of serviceprovision and gives an idea of coverage in ANCsettings where PMTCT interventions areavailable. It does not attempt to inform serviceproviders about the points in the counseling andtesting cycle at which women drop outIt is important that program managers employ aseries of lower-level indicators for determininglosses to follow up. Because the quality of servicesis not being measured, information on drop-outsand the points at which they occur is of limiteduse if not followed up with operations researchaimed at discovering why women are failing tocomplete the cycle.36Chapter 2

OF HIV-P-PPERERCENTCENTAGEOF-POSITIVEPREGNREGNANTWOMENRECEIVINECEIVINGA COMPLETECOURSEOF ARV PROPHYLOPHYLAXISTO REDUCEMTCTIN ACCCCORDORDANANCEWITH A NATIONTIONALLALLY APPRPPROVEDTREAREATMENTPROTOCOCOLOL (ORWHO/UNAIDSSTANDANDARDSARDS) IN THE PRECEDIN12 MONTHSRECEDING 12 MUNGASS Core e Outcome IndicatorDefinitionThe percentage of HIV-positive pregnant womenreceiving a complete course of ARV prophylaxisto reduce MTCT in accordance with a nationallyapproved treatment protocol (or WHO/UNAIDSstandards) in the preceding 12 months.Numerator: Number of HIV-positive pregnantwomen receiving a complete course of ARVprophylaxis to reduce MTCT in accordancewith a nationally approved treatment protocol(or WHO/UNAIDS standards) in thepreceding 12 months.Denominator: The estimated number of HIVinfectedpregnant women giving birth in thepreceding 12 months.The definition of a “complete course of ARVprophylaxis” will depend on the country’s policyon ARV prophylaxis to reduce the risk of MTCTand may or may not include a dose for newborns.Details of the definition used should be provided.Measurement ToolsFacility-based national MIS; program monitoring;HIV sentinel surveillance-based records; NationalStatistical Office estimatesA tool for the measurement of this indicator isprovided in Monitoring the Declaration ofCommitment on HIV/AIDS: Guidelines onConstruction of Core Indicators.What It MeasuresThis is an UNGASS national program andbehavior indicator. It assesses progress inpreventing mother-to-child HIV transmissionthrough the provision of ARV prophylaxis.How to Measure ItThe number of HIV-infected pregnant womenwho have been provided with ARV prophylaxisduring the preceding 12 months in order to reducethe risk of MTCT is obtained from programmonitoring records. Only those women whocompleted the full course should be included inthe numerator. The denominator represents theestimated number of women in need of ARV, thatis, the number of HIV-infected women to whomARV prophylaxis to reduce the risk of MTCTcould potentially have been given. Thedenominator is estimated by multiplying the totalnumber of women who gave birth in the preceding12 months (central statistics office estimates ofbirths) by the most recent national estimate ofHIV prevalence in pregnant women (HIV sentinelsurveillance antenatal clinic estimates).Whether women who receive ARV prophylaxisfrom the private sector and NGO clinics shouldbe included in the calculation of the indicator isleft to the discretion of the country concerned.This decision should be based on a frank appraisalof how often ARV for pregnant HIV-infectedwomen is provided outside the government sector,and should be noted and applied consistently incalculating both the numerator and denominator.Private sector and NGO clinics that provideprescriptions for ARVs but assume that the drugswill be acquired elsewhere by the individuals arenot included in this indicator, even though suchclinics may be the major providers of services forthe reduction of MTCT. The key feature is theactual provision of the drugs.This indicator should be measured every two tothree years.Prevention of Mother-to-Child Transmission (PMTCT) of HIV37

Strengths and LimitationsThis indicator has the following weaknesses:(1) ANC data are often incomplete and may notreflect the true situation.(2) There may be selection bias because only thosewomen are included who self-select to accessservices.(3) Every country has its own definition of a fullcourse of ARV treatment.(4) The indicator does not assess treatmentcompliance, and, as currently defined,measures need. It does not assess whatpercentage of women accessing ANC serviceswhere PMTCT services are available actuallyavail themselves of the intervention.(5) As the number of women provided withHAART increases over time, the need forspecific ARV distribution to prevent verticaltransmission may lessen. It will be necessaryto develop specific indicators in order tocapture information on this matter.38Chapter 2

OF HIV-I-IPERERCENTCENTAGEOF-INFECTEDINFNFANTANTS BORNTO-INFECTEDMOTHERSHIV-I-IUNGASS Core e Impact IndicatorDefinitionThe percentage of HIV-infected infants born toHIV-infected women.Measurement ToolsFacility-based MIS (national counts of pregnantwomen receiving ARV to prevent MTCT, as pernational guidelines); background rate of MTCTwithout PMTCT intervention; current estimatesof efficacy of national PMTCT drug regimen(average)The complete tool for measuring this indicatorcan be found in Monitoring the Declaration ofCommitment on HIV/AIDS: Guidelines onConstruction of Core Indicators (UNAIDS, 2002).What It MeasuresThis UNGASS indicator measures the impact onMTCT reduction of the provision of ARVs topregnant women in order to prevent verticaltransmission. The UNGASS targets are a 20%reduction in the percentage of HIV-infectedinfants born to HIV-infected mothers by 2005 anda 50% reduction by 2010.How to Measure ItThis indicator is measured by taking the weightedaverage of the probabilities of MTCT for pregnantwomen receiving and not receiving ART, theweights being the proportions of women receivingand not receiving ARV, respectively.The indicator is calculated using the followingformula:where:Indicator score = {T*(1 - e) + (1 - T)} * vT = proportion of HIV-infected pregnant womengiven ART (this is the proportion obtained in theUNGASS indicator on ARV prophylaxis)v = MTCT rate in the absence of treatmente = efficacy of treatment providedDefault values of 0.25 and 0.50, respectively, canbe used for v and e. However, if scientific estimatesof the efficacy of the specific forms of treatment(i.e. combination therapies) employed in thecountry are available, these should be used, andthis should be noted in the calculations.This indicator should be measured every two tothree years.Strengths and LimitationsIf an infant becomes positive, the indicator cannotdistinguish between different causes of infection,i.e. treatment failure or infection duringbreastfeeding. The indicator may thereforeunderestimate the rates of MTCT in countrieswhere long periods of breastfeeding are common.Conversely, rates may be overestimated incountries where other MTCT preventioninterventions are common, e.g. cesarean section.The proportion of HIV-infected pregnant womengiven treatment, T, may be a poor estimate in placeswhere the usage of ANC clinic services is low.Prevention of Mother-to-Child Transmission (PMTCT) of HIV 39

RefererencesencesDe Cock KM, Fowler MG, Mercier E, Vincenzi I, Saba J, Hoff E, Alnwick DJ, Rogers M, & Shaffer N.(2000). Prevention of Mother-to-Child HIV Transmission in Resource-Poor Countries: Translating Researchinto Policy and Practice. Journal of the American Medical Association, 283,9:1175-1182.UNAIDS. (1996). Care and Support Indicators: Protocol for the Evaluation of Case Management of HIV/AIDSPatients. UNAIDS/96.3. Geneva, Switzerland: UNAIDS. Retrieved from http://www.cpc.unc.edu/measure/guide/tools/whocare/unaidscare.pdfUNAIDS. (2000a ). National AIDS Programmes: A Guide to Monitoring and Evaluation. Geneva, Switzerland:UNAIDS. Retrieved from http://www.unaids.org/publications/documents/epidemiology/surveillance/JC427-Mon&Ev-Full-E.pdfUNAIDS. (2000b). Local Monitoring and Evaluation of the Integrated Prevention of Mother to Child HIVTransmission in Low-income Countries. Geneva, Switzerland: UNAIDS. Retrieved from http://www.unaids.org/publications/documents/mtct/ME2000.docUNAIDS. (2000c). Tools for Evaluating HIV Voluntary Counseling and Testing. UNAIDS/00.09E. Geneva,Switzerland: UNAIDS. Retrieved from http://www.un.org/Depts/dpko/medical/pdfs/602unaidscounsel.pdfUNAIDS. (2002). Monitoring the Declaration of Commitment on HIV/AIDS: Guideline on Construction ofCore Indicators. Geneva, Switzerland: UNAIDS. Retrieved from http://www.unaids.org/UNGASS/docs/JC718-CoreIndic_en.pdfUNAIDS. (2004). Global Summary of the AIDS Epidemic: December 2004. Geneva, Switzerland: UNAIDS.Retrieved from http://www.unaids.org/wad2004/report_pdf.htmlUNGASS. (2001). Declaration of Commitment on HIV/AIDS: Global Crisis - Global Action. United NationsGeneral Assembly Twenty-Sixth Special Session. Retrieved from http://www.un.org/ga/aids/coverage/FinalDeclarationHIVAIDS.htmlWHO. (2000). Protocol for the Evaluation of HIV/AIDS Care and Support. [Draft]. Geneva, Switzerland:World Health Organization. Retrieved from http://www.emro.who.int/gfatm/guide/tools/whocare/whocare.pdfWHO. (2004). National Guide to Monitoring and Evaluating Programmes for the Prevention of HIV in Infantsand Young Children. Geneva, Switzerland: World Health Organization.Chapter 2. Prevention of Mother-to-Child Transmission (PMTCT) of HIV41

3NEWBORNHEALEALTHIndicators:• Number of health facilities providing basic and comprehensive emergencyobstetric care functions per 500,000 population• Proportion of hospitals and maternity facilities designated as baby friendly• Proportion of health workers who are competent in neonatal resuscitationupon completion of training• Proportion of pregnant women attending antenatal clinics who arescreened for syphilis• Proportion of babies who receive eye prophylaxis care within one hour of birth• Percentage of HIV-positive pregnant women receiving a completecourse of ARV prophylaxis to reduce MTCT in accordance with anationally approved treatment protocol (or WHO/UNAIDS standards)in the preceding 12 months (cross-referenced in Chapter two)• Percent of pregnant women who received at least two antenatal care visits• Proportion of pregnant women receiving at least two doses of tetanus-toxoidvaccine• Proportion of pregnant women receiving intermittent preventive treatmentor malaria prophylaxis, according to national policy• Proportion of pregnant women who know two or more newborn danger signs• Proportion of deliveries occurring in a health facility• Proportion of deliveries with a skilled attendant at birth• Maternal mortality ratio• Proportion of newborns who receive thermal protection immediately afterbirth• Timely initiation of breastfeeding (cross-referenced in Chapter eight)• Exclusive breastfeeding rate (cross-referenced in Chapter eight)• Proportion of women who receive two high-dose supplements of vitamin Awithin six weeks of giving birth• Preterm birth rate• Proportion of live births with low birth weight• Late fetal death rate (cross-referenced in Chapter nine)• Perinatal mortality rate (cross-referenced in Chapter nine)• Cause-specific perinatal mortality rate• Birth weight specific mortality rate (cross-referenced in Chapter nine)• Number of neonatal tetanus cases• Neonatal mortality rate (cross-referenced in Chapter nine)

HAPTER 3. NCHAPTER3. NEWBEWBORNHEALEALTHThe goal of many programs in developingcountries is to improve maternal andnewborn health and survival. Until recently,however, newborn health was relatively neglectedin both the international child health and safemotherhood movements, and few programsfocused specifically on improving newbornsurvival. A prime reason that newborn health hasreceived such low priority is the general lack ofawareness of the sheer numbers of early infantdeaths. WHO estimates that each year more than8 million infants die in the first year; of these,almost two thirds (5.1 million) die in the firstmonth, and of these, two thirds die within the firstday (Lawn, McCarthy & Ross, 2001). Virtuallyall of these deaths occur in developing countries.Although post-neonatal mortality has declinedsubstantially, neonatal deaths have declined onlyslightly, thereby representing a growing proportionof overall infant deaths (Espeut, 1998).A second factor has been the perception thatsophisticated technologies are required tosignificantly reduce perinatal and neonatalmortality. On the contrary, most newborn deathsin developing countries can be prevented byinterventions already widely used. The mostcommon causes of neonatal mortality – infections,asphyxia and birth injuries – can be prevented bysimple cost-effective interventions that also benefitthe mother. These interventions include antenatalmalaria prevention and treatment, tetanus toxoidimmunization, the detection and management ofsexually transmitted infections, and access to aclean and safe delivery (WHO, 1996a).Furthermore, providing all infants with an“essential package of newborn care” (see Table 3.1)including appropriate resuscitation, warmth,cleanliness and hygiene, clean cord care, and earlyexclusive breastfeeding also increases survival andreduces the proportions of surviving infants withdisability (WHO, 1996c; WHO, 2001a).Chapter 3. Newborn HealthCompared to other programmatic (technicalintervention) areas discussed in this guide,newborn health is one of the least developed.Systematic review of operations research studieshave identified which interventions are likely toeffectively reduce newborn mortality, but howthese services should be scaled up, by whom, andat what cost must still be determined. Themonitoring and evaluation of these programs isalso in its infancy, and many new data-gatheringtools, analytical approaches, and indicators needto be developed and tested. Because of the closelink between maternal and newborn health,however, many output indicators appropriate fornewborn health are used extensively in safemotherhood programs. Indeed, separatingnewborn health indicators from those pertainingto maternal health may create a false dichotomywhen the antecedents of a poor pregnancyoutcome overlap with the program interventionsrequired to address them.However, our purpose in having a specific sectionon newborn health is to acknowledge growingawareness of the importance of newborn healthand to highlight the fact that despite the manyparallels between maternal and newborn healthprograms, important differences influence the waythat programs are monitored and evaluated. Thesedifferences arise, not only because programinterventions may vary, but because interventionsthat benefit both mothers and babies maydifferentially affect mortality. For example, someindicators strongly associated with maternalsurvival such as antenatal care and skilledattendance at birth may not have an equally strongassociation with perinatal survival. Otherinterventions such as immunizing pregnantmothers against tetanus are more likely to bemonitored in newborn health programs than insafe motherhood programs.43

Table 3.1. WHO essential newborn care package1) Cleanliness: clean delivery and clean cord care for the prevention of newborn infections(tetanus and sepsis)2) Thermal protection: prevention and/or management of neonatal hypothermia andhyperthermia3) Early and exclusive breastfeeding4) Initiation of breathing, resuscitation5) Eye care: prevention and management of ophthalmia neonatorum6) Immunization (BCG, Oral polio, Hepatitis B)7) Management of newborn illness8) Care of the preterm and/or low birth weight newbornSource: WHO (1996a).Methodological Challenges to EvaluatingNewborn Health ProgramsSome of the challenges of evaluating newbornhealth programs include the following:Countries define births, deaths, and “newborn period”in different ways, making valid internationalcomparisons difficult.Meaningful use of any indicator is only feasiblewhen standard definitions are used and applied.The lack of a generally agreed-upon definition ofthe “newborn period” may limit comparisonsacross countries and programs. In some settings,“newborn” may refer to infants up to a few days ofage and in other settings to infants up to severalweeks of age. In this guide, the term “newborn”refers to the neonatal period (i.e., the first 27completed days of life).Outcomes need to be measured for two individuals,the mother and the baby.Newborn health programs (like safe motherhoodprograms) need to consider the outcomes for twoindividuals: the mother and the baby. Just becausethe newborn receives a postnatal checkup, it doesn’tmean the mother receives one also.Interpreting whether outcomes are attributable toprogram interventions is difficult because mostinterventions consist of “bundled” services.Demonstrating change due to a newborn healthprogram is difficult because programs usuallyprovide a package of care to communities ratherthan a single intervention. Therefore, suchprograms do not lend themselves easily to twocommon experimental designs: randomizedcontrol trials and cluster randomized communitytrials. Many programs adopt “before-after”designs for evaluation purposes that candemonstrate “plausible association” but that fallshort of causality (UNFPA et al., 1997).Measuring perinatal and neonatal morbidity is verydifficult.Estimates of newborn morbidity are important fordesigning effective program interventions. Aswith safe motherhood, however, existing estimatesof newborn morbidity are usually derived fromfacility data and are unlikely to reflect the trueburden of morbidity in the community where useof health facilities is low. Although communitymembers can learn to diagnose illness in a sicknewborn (Bang et al., 1999), illness is oftendifficult to recognize because babies usually44 Chapter 3

present with relatively non-specific symptoms,such as poor feeding and lethargy. Few facilitieshave adequate diagnostic facilities when ill babiesdo eventually present for care. Assigning a causeof death may be difficult because many differentdiseases may present with the same symptoms, andmany babies die at home before ever reachingmedical attention.New program indicators are required at theindividual, community, and facility level.Much of the discussion on the challenges ofmonitoring and evaluating newborn health has sofar focused on newborn mortality because of therelative lack of experience with output indicatorsfor newborn health. Although mortality indicatorsclearly have their place and provide the only directmeasure of the long-term objective of mostprograms, output indicators need to be developedto measure the wide range of interventionsrequired to improve newborn health and survival.Output indicators are required for measuring theavailability, accessibility, quality, utilization anddemand for services at the facility level and foroutreach services where the provision of newbornhealth services has historically been overlooked.In addition, intermediate outcome indicators arerequired for monitoring and evaluatinginterventions at the individual and communitylevels. Many infants become ill and die beforeever reaching medical care. It is particularlyimportant to develop indicators that helpprograms understand community knowledge,attitudes, and behaviors in response to newbornillness and to determine which interventions arethe most effective.Relationship Between Maternal and NewbornHealthFigure 3.1 illustrates the links between maternaland newborn health from before pregnancy to afterdelivery. The figure does not address some of thesystem-level determinants – the social, cultural,economic, political, and legal factors that influencematernal and newborn health. The primarypurpose of the figure is to show where maternalhealth interventions can promote and improve thehealth status of the newborn, as well as the levels(family, community, and services) at which theimpact of these interventions should be measured.Figure 3.1. Maternal & newborn health linkages.Newborn Health45

As depicted in Figure 3.1, several interventionsneed to be implemented during pregnancy toincrease newborn survival. These include highquality antenatal care, timely recognition andmanagement of obstetric complications, goodnutrition, and micronutrient supplementation(including iron and folate supplementation whereanemia is common; vitamin A supplementationwhere vitamin A deficiency is prevalent; andiodization of salt and treatment of iodinedeficiency with iodized oil).As infections during pregnancy can have a seriouseffect on the survival of the newborn, newbornhealth requires the prevention and treatment ofinfections in pregnancy. Relevant interventionsinclude the presumptive treatment of malaria andhookworm in endemic areas, identification andtreatment of syphilis, and tetanus toxoidimmunization. Another important interventionis the promotion of voluntary counseling andtesting for HIV/AIDS for mothers to reduce therisk of mother-to-child transmission of HIV/AIDS. Promoting newborn health also includesmobilizing facilities, providers, communities, andfamilies around birth preparedness andcomplication readiness for both the mother andthe newborn.The Selection of IndicatorsMost indicators in this section of the guide areintended for use at the national level or in thecontext of large-scale programs, but many can beused in a much wider monitoring and evaluationcontext. Indicators were selected on the basis thatthey:• Are widely used by international organizationsor ministries of health;• Have a relatively strong link to health ormortality outcomes; and• Will likely provide valid comparisons at thenational and international level.The indicators included in this section of the guiderelate directly to safe motherhood, newborn health,and newborn health care and include thefollowing, which are based on WHO and CDCrecommendations for monitoring and evaluatingnewborn health at the global and national levels.Output Indicators• Number of health facilities providing basic andcomprehensive emergency obstetric carefunctions per 500,000 population• Proportion of hospitals and maternity facilitiesdesignated as baby friendly• Proportion of health workers who arecompetent in neonatal resuscitation uponcompletion of training• Proportion of pregnant women attendingantenatal clinics who are screened for syphilis• Proportion of babies who receive eyeprophylaxis care within one hour of birth• Percentage of HIV-positive pregnantwomen receiving a complete course of ARVprophylaxis to reduce MTCT in accordancewith a nationally approved treatmentprotocol (or WHO/UNAIDS standards)in the preceding 12 monthsOutcome Indicators• Percent of pregnant women who received atleast two antenatal care visits• Proportion of pregnant women receiving atleast two doses of tetanus-toxoid vaccine• Proportion of pregnant women receivingintermittent preventive treatment or malariaprophylaxis, according to national policy• Proportion of pregnant women who know twoor more newborn danger signs• Proportion of deliveries occurring in a healthfacility• Proportion of deliveries with a skilledattendant at birth• Maternal mortality ratio46 Chapter 3

• Proportion of newborns who receive thermalprotection immediately after birth• Timely initiation of breastfeeding• Exclusive breastfeeding rate• Proportion of women who receive two highdosesupplements of vitamin A within sixweeks of giving birthImpact Indicators• Preterm birth rate• Proportion of live births with low birth weight• Late fetal death rate• Perinatal mortality rate• Cause-specific perinatal mortality rate• Birth weight specific mortality rate• Number of neonatal tetanus cases• Neonatal mortality rateWe are aware that some programs cannot measurethe neonatal health outcomes in the guide, andfor these programs, our selection will be less useful.Certain indicators (for example, Proportion ofPregnant Women Who Know Two or MoreNewborn Danger Signs) have been field testedby some groups, but not widely adopted. Thisindicator is included because of the need tostimulate debate and discussion on appropriatenewborn health process indicators, even thoughwe recognize that this indicator may not meet allthe criteria for a good indicator (WHO, 1997).Some indicators (for example, Proportion ofBabies Who Receive Eye Prophylaxis CareWithin One Hour of Birth) are difficult tomeasure but have been included nonethelessbecause they are a critical component of the WHOEssential Newborn Care Package. In the next fewyears, as awareness of the problem of newbornmortality grows, no doubt those working innewborn health will move toward a consensus onthe indicators appropriate for monitoringnational-level programs.Newborn Health47

NUMBEROF HEALEALTHFACILITIESPROVIDINVIDING BASICANDCOMPREHENSIVEEMERMERGENGENCYOBSBSTETRICCAREPEROPULATION500,000 POPULDefinitionNumber of health facilities providing basic andcomprehensive emergency obstetric care functionsper 500,000 population.Numerator: Number of facilities providing allstandardized basic and comprehensiveemergency obstetric care functions.Denominator: Total population of catchmentarea.“Emergency obstetric care” functions are defined as:• Administration of parenteral antibiotics;*• Administration of parenteral oxytocic drugs;• Administration of parenteral anticonvulsants forpre-eclampsia and eclampsia;• Performance of manual removal of placenta;• Performance of removal of retained products(e.g., manual vacuum aspiration);• Performance of assisted vaginal delivery(e.g., vacuum extraction, forceps);• Performance of surgery (e.g. cesarean section);and• Performance of blood transfusion.Health facilities are divided into those that provide“basic” emergency obstetric care (EmOC) and“comprehensive” EmOC. If a facility has performedeach of the first six functions in the past three months,it qualifies as providing basic EmOC. If it hasprovided all eight of the functions in the past threemonths, it qualifies as a “comprehensive” EmOCfacility (UNICEF/WHO/UNFPA, 1997). Note* Parenteral administration of drugs means by injection orintravenous infusion (“drip”).Newborn Healththat there has been confusion with changingterminology from “essential” to “emergency” obstetriccare. For the purpose of this manual, we will refer toemergency obstetric care.Measurement ToolsHealth facility assessments that examine medicalrecords or service statistics; personal interviews withknowledgeable staff who attend obstetric patients(These are a second, albeit, potentially more biasedsource of information than written records.)What It MeasuresThis indicator measures the availability of life-savingobstetric care services. Due to the difficulties ofmeasuring maternal mortality and morbidity, a seriesof process indicators are being suggested as analternative (Goodburn, 2002). This indicator is oneof those suggested indicators. It distinguishesbetween “basic” and “comprehensive” emergencyobstetric care services to emphasize that maternallives can be saved not only in hospitals providing allthe services listed above, but also at health centers orsmaller hospitals with basic services.How to Measure ItThis indicator is calculated by counting the numberof facilities meeting the requirements for “Basic” (or“Comprehensive”) EmOC, dividing by the totalpopulation of the catchment area, and multiplyingthe result by 500,000. Civil registration andpopulation censuses provide information for thedenominator.The indicator should be calculated separately to showthe availability of Basic EmOC services and theavailability of Comprehensive EmOC services. Onlyfacilities currently providing all the signal functionsin either the Basic or Comprehensive EmergencyObstetric Care lists should be included in the49

numerator. Ideally, the facility should have all thesignal functions available 24 hours a day and sevendays a week.UNICEF/WHO/UNFPA (1997) recommends aminimum acceptable level of at least four basic andone comprehensive EmOC facilities per 500,000population. This indicator can be calculated forsmaller geographic areas to show the distribution ofEOC facilities at a sub-national level.Strengths and LimitationsThis indicator is relatively easy to produce andshould respond to changes within a fairly shortperiod of time (e.g., 6-12 months). Note that thelist of signal functions is intentionally brief tofacilitate assessment and monitoring; it does notconstitute the complete list of services that eithera Basic or Comprehensive EmOC facility shouldprovide. Valuable services are omitted in thedefinition of EmOC facility. For example, use ofanesthesia is not included, although it is assumednecessary for obstetric surgery.Generally, this indicator applies to a large regionor country and tells us whether there are problemsin the availability of EmOC services in the generalpopulation. However, it does not tell us wherefacilities are needed or why existing EmOCfacilities are not being used. The indicator is notnecessarily a reflection of accessibility of facilitiesbecause it contains no information on theirgeographical distribution, referral systems,transport, cultural, and economic accessibility, orthe uptake of care. Furthermore, this indicatormay not reflect true differences in the availabilityto the population in need of EmOC (i.e. pregnantwomen) where there are differences in theproportion of women of reproductive age in thepopulation and their fertility rates.Generally, facility-based assessments should coverall the facilities in a specific area. However, privatefacilities may be under-represented compared topublic facilities. Although geographicallyrepresentative samples of facilities are possible inhealth facility surveys, such as in the Service50Provision Assessment (SPA), these surveys maynot always include all the signal functions listedpreviously (MEASURE DHS+, 2000).If areas fall short of the minimum level mentionedpreviously, they may upgrade existing facilitiesand/or build new ones. If the minimum level ismet, evaluators should study the geographicaldistribution by looking at smaller divisions of thepopulation. National summary measures may hideimportant sub-national disparities; hence,disaggregation by geographic (urban/rural) and byadministrative (public/private) divisions isrecommended (Bertrand and Tsui, 1995).The use of this indicator in a wide variety ofcountries has brought to attention at least threedifficulties in its application. First, wheregeographical terrain is particularly challenging andtransportation is precarious (such as in themountains of Nepal and Bhutan), the ratio offacilities to the population may require adjustmentfor local use. Second, the reference period forassessing whether a signal function or procedurehas been performed is generally three months, butwhen patient volume is low, one or more of thesignal functions may not be performed becausean occasion did not present itself, not for lack ofinfrastructure or provider skills. Finally, a thirdsituation concerns normative medical practice thatfails to include one of the procedures, for example,assisted vaginal delivery. In some countries,vacuum extraction or a forceps delivery is no longertaught to medical students or midwives, and onlya few older providers are experienced at performingthese procedures.In attempting to solve these problems, one mayconsider modifying the indicator in several ways.However, researchers must documentmodifications made to the definition and thecalculation of the indicator in order to informcomparisons of facilities across time and space. Ifcountry-specific criteria were established in thedefinition of Basic or Comprehensive EmOC, orif a particular signal function was omitted fromthe definition, these changes should bedocumented as well.Chapter 3

PROPOROPORTIONOF HOSPITOSPITALALS AND MATERNITTERNITY FACILITIESDESIGNESIGNATEDAS BABABY FRIENDLRIENDLYDefinitionThe proportion of hospitals and maternityfacilities that have been accredited as “BabyFriendly” according to the ten UNICEF/WHOcriteria related to breastfeeding and newborn care.Numerator: Number of hospitals andmaternity facilities accredited as “BabyFriendly.”Denominator: Total number of hospitals andmaternity facilities that handle deliveries.To be designated as “Baby Friendly,” the hospitalmust:• Have a written breastfeeding policy that isroutinely communicated to all health care staff;• Train all health-care staff in the skills necessaryto implement this policy;• Inform all pregnant women about the benefitsand management of breastfeeding;• Help mothers initiate breastfeeding within anhour of birth;• Show mothers how to breastfeed and how tomaintain lactation, even if they should beseparated from their infants;• Give newborn infants no food or drink otherthan breast milk, unless medically indicated;• Practice “rooming in” by allowing mothers andinfants to remain together 24 hours a day;• Encourage breastfeeding on demand;• Give no artificial teats, pacifiers, dummies, orsoothers to breastfeeding infants; and• Foster the establishment of breastfeedingsupport groups and refer mothers to them ondischarge from the hospital or birthing center.Measurement ToolsUNICEF/WHO/Wellstart Baby FriendlyHospitals Initiative (BFHI) internal selfassessmentand external evaluation instrumentsWhat It MeasuresThis indicator provides useful information on theavailability of baby-friendly services in a givencountry. The BFHI is a joint UNICEF/WHO/Wellstart initiative aimed at increasingbreastfeeding rates and encouraging globalstandards for maternity services in hospitals andmaternities.How to Measure ItData requirements are the number of maternitiesmeeting BFHI criteria and the total number ofmaternities and hospitals. Facilities first conducta self-assessment, then independent assessorsappointed by the national BFHI committee orUNICEF country offices evaluate them accordingto the criteria mentioned in the previous column.These same bodies aggregate information on thenumbers and proportions of facilities acquiring“Baby Friendly” status for national and globalreporting (WHO, UNICEF, and WellstartInternational, 1999).Strengths and LimitationsThe number of facilities achieving “Baby Friendly”status may be presented more often than theproportion because of difficulties in ascertainingthe total number of maternities required for thedenominator. Ascertaining the number ofmaternities in the private sector is particularlydifficult, and in many cases, private facilities maynot be represented in national estimates. Thenumber of facilities achieving “Baby Friendly”status is of limited use for regional and crosscountrycomparisons because it is clearly affectedNewborn Health51

y geographic size. For example, by December2000, 6312 hospitals in China (or 47% of alleligible facilities) had achieved “Baby Friendly”status compared to 232 (or 66% of all eligiblefacilities) in Kenya.Second, the listing of facilities that are recordedas “Baby Friendly” may be out of date becauseperiodic reaccreditation to maintain standards isvoluntary and depends on the interest andmotivation of each individual facility. The date ofacquiring “Baby Friendly” status and whetherreaccreditation has occurred are not routinelyrecorded.52Chapter 3

PROPOROPORTIONOF HEALEALTHWORKERSWHOARECOMPETENTIN NEONEONATALAL RESUSCITESUSCITATIONTION UPONCOMPLETIONOF TRAININRAININGDefinitionProportion of health workers who are competentin neonatal resuscitation upon completion oftraining.Numerator: Number of health workers whoare competent in neonatal resuscitation uponcompletion of training.Denominator: Total number of health workerstrained.The definition of “competency” depends on nationaltraining objectives. Usually, “competent” refers tothe fact that the trainee can deliver the serviceaccording to a set standard.Measurement ToolsCompetency test (often in the form of a checklistadministered by the trainers and/or external expertobserver)What It MeasuresThis indicator measures competency in neonatalresuscitation among health workers who havecompleted relevant training. It reflects both theadequacy of the training with respect to these skillsand the ability of trainees to absorb theinformation. Resuscitation is needed when anewborn suffers from birth asphyxia. Birthasphyxia is defined as the failure to initiate andsustain breathing at birth. WHO estimates thatapproximately one to five percent of all newbornswill require resuscitation at birth which accountsfor up to 6 million babies per year. Of these aboutone million will die, and an unaccounted numberwill suffer from long-term disabilities (WHO,1998d).The incidence of birth asphyxia is higher indeveloping countries because of higher prevalenceof risk factors such as poor health of women whenthey become pregnant; higher incidence ofpregnancy and delivery; inadequate or nonexistentcare during delivery and labor; and highincidence of pre-term delivery. An internationalstudy found that 80% of babies requiringresuscitation needed only a bag and mask and roomair (Saugstad, Rootwelt, and Aalen, 1998). Moreadvance complex technologies are not alwaysnecessary.How to Measure ItMeasurement of this indicator requires that healthworkers undergo competency-based training forneonatal resuscitation. Four main pieces ofinformation are needed to calculate this indicator:• Training records• Written tests (e.g., pre- and post-tests ofknowledge)• Results of observer checklists, pre and posttestsof skills• National or institutional standards for trainingand service deliveryCompetency should be assessed after training hasbeen completed using a model. The instrumentused in the evaluation is a standardized checklistincluding the relevant skills and steps in newbornresuscitation. Each step of the skill is scored bythe evaluator to indicate if the skill was donecorrectly (2 points), done partly correctly (1 point),or done incorrectly or not at all (0 points). Table3.2 summarizes different components of neonatalresuscitation.Newborn Health53

The competency assessment may require thehealth worker to do the following:(a) Explain while demonstrating the first fivesteps of newborn resuscitation;(b) Explain while demonstrating the two thingsthat are evaluated in a baby to decide if moreresuscitation is needed;(c) Explain while demonstrating how to doresuscitation including infection preventionsteps;(d) Explain what to do if the chest does not riseafter the baby is given the first breath;(e) Explain while demonstrating how to keep thebaby warm and stimulated when the baby isbreathing but the color is blue.The strength of this indicator is that competencyin neonatal resuscitation can be assessed every sixmonths (as resources permit) against WHO ornational standards. When measuring providercompetency at a later date (using directobservation), the indicator is useful in determiningthe retention of skills acquired during training,and for identifying possible candidates forretraining.This indicator can also be measured through directobservation of health worker practices at healthfacilities with high client volume. The same skillschecklists used during the training can be used toassess on-the-job performance. A score of at least80 percent can be used to represent “competency”for skills and a score of at least 70% to represent“competency” for knowledge. Low scores, withno critical steps missing, may reflect inadequaciesin the course and/or in the ability of theparticipants to absorb the information.Strengths and LimitationsThe limitation of this indicator is that it would bedifficult to compare the results from this indicatoracross countries or even across programs within agiven country. Learning objectives, courses, andevaluation tools are not typically standardized. Atthe field level, there are inconsistencies in termsof the criteria used to define competency. Someprograms would expect a 100% grade before thetrainee would be judged competent on a batteryof skills, whereas another program might considera person competent if only 70% of the tasks arecorrectly completed. In some cases, local standardsfor neonatal resuscitation may not exist, in whichcase international standards can be used.54Chapter 3

Table 3.2. Essential newborn care intervention sub-packagePACKAGE COMPONENT-NEONATAL RESUSCITATIONIdentification of babies to resuscitate:X Not breathingX Blue color of mouth and body, or floppy and whiteStandard and staff:X Develop resuscitation standards for different levels at your settingX Use existing national or WHO guidelines as a basisX Provide competency-based training for all staff who will be at the deliveriesX Provide supervision and ongoing training, as these skills can be lostSupplies and equipment:X Dry clean clothX Bag and mask (ambu bag)X Suction apparatus (a range of options are available in WHO guidelines)X GlovesAdditional equipment:X Shelf to put the baby onX Method to keep the baby warm, such as overhead light bulbsX Oxygen supplyImmediate Actions:X Dry the baby and cover with a clean clothX Suction the mouth and nose if required (over-suctioning at the back of the throat canmake the baby's condition worse)X Place the baby correctly, with a small roll of cloth under the neck to extend it slightlyX Place the mask (attached to the bag) firmly over the baby's mouth and nose and form asealX Squeeze the bag to inflate the lungs at a rate of 40 respirations per minuteX Watch the baby's chest carefully to see that the chest is rising and falling as you squeeze thebagX If the baby starts breathing regularly, stop using the ambu bagX If there is no gasping or breathing at all, stop resuscitating after 20 minutesX If there was gasping but no spontaneous breathing, stop resuscitation after 30 minutesX Universal precautions should be observed, including hand washing, disinfecting allequipment, and use of gloves, if availableNewborn Health55

Table 3.2. Essential newborn care intervention sub-package (continued)Mouth to mouth resuscitation:If a bag and mask are not available, mouth to mouth resuscitation is effective. The risk forinfection to the resuscitator can be reduced by cleaning blood and mucus from the baby's faceand mouth with a cloth and placing the cloth over the baby's mouth and nose before startingto ventilate. Seal the mouth and nose with your mouth and blow at a rate of 40 respirations perminute.Source: WHO/RHT/SMS/98.1.56Chapter 3

PROPOROPORTIONOF PREGNREGNANTWOMENATTENDINTTENDINGANTENNTENATALAL CLINICSWHOARESCREENEDFORSYPHILISDefinitionProportion of pregnant women attendingantenatal clinics who are screened for syphilis.Numerator: Number of pregnant womenattending antenatal clinics who are screenedfor syphilis.Denominator: Total number of pregnantwomen attending antenatal clinics.Measurement ToolsClinic registries (data on first visit) or individualprenatal records (individual ANC records/cardsafter births or immediately postpartum); healthfacility surveysWhat It MeasuresSyphilis infection is a major cause of maternalmorbidity and perinatal morbidity and mortalityin the developing world. For many Africancountries, reported prevalence of syphilis amongpregnant women at sentinel surveillance sitesranges between 10-15 percent, with over half ofthese pregnancies resulting in an adverse outcome,such as abortion, stillbirth, low birth weight,premature delivery, or congenital infection(WHO, 1991b). Because adverse outcomes fromsyphilis are preventable, and screening andtreatment in pregnancy are highly cost effective,many countries have adopted universal syphilisscreening for pregnant women as a national policy(Gloyd et al., 2001).This indicator measures the extent to which ANCclients are screened for syphilis. Since all womenattending for ANC should be screened for syphilisat least once during pregnancy, this measure canalso potentially serve as a proxy measure of thequality of antenatal care services (UNFPA, 1998a).Newborn HealthFurthermore, when an explicit standard exists thatall women should be tested at least once duringpregnancy, the indicator may also be used as abenchmark to audit provider (or system)performance against compliance with localscreening policy.How to Measure ItResearchers may routinely collect data to calculatethis indicator if antenatal clinic registries recordcompleted syphilis screening. Most often,however, the information is collected in the contextof special surveys that review the antenatal cliniccards of women who have had a recent birth.Researchers may conduct these surveys in facilitiesor in the community, if women keep their antenatalcards. Health facility exit interviews and providerobservations (MEASURE DHS+, 2001; WHO,1998a) may provide a baseline measure forevaluation purposes, but are limited because theyassess women who have not yet completedantenatal care and who theoretically could still betested (MEASURE DHS+, 2001; WHO, 1998a).Strengths and LimitationsThis indicator is a facility-based measure and doesnot represent the general population, particularlywhen ANC coverage is low. In addition, wherethe indicator is obtained by record review, thevalidity of the findings depends on the quality andcompleteness of the data. Incomplete datarecording may further indicate low service quality.Adequate syphilis screening does not equate withadequate syphilis treatment because studies showthat despite effective screening, inadequatetreatment may be an important cause ofpreventable perinatal death. In high prevalenceareas, even when syphilis testing is theoreticallyuniversal, most women were not tested (Gloyd etal., 2001).57

PROPOROPORTIONOF BABIESWHORECEIVEEYEPROPHYLOPHYLAXISCAREWITHINONEHOUROF BIRIRTHDefinitionProportion of babies who receive eye prophylaxiscare within one hour of birth in a specified period.Numerator: Number of babies who receive eyeprophylaxis care within one hour of birth in aspecified period.Denominator: Total number of live births inthe same period.“Eye prophylaxis” involves cleaning the eyes afterbirth and applying either silver nitrate drops (1%),tetracycline ointment (1%), or erythromycinointment (0.5%) within the first hour of birth(WHO, 1996a; Lawn et al., 2001). The type ofmedication used depends on the localepidemiological situation.Measurement ToolsHealth facility survey (direct observation ofdeliveries)What It MeasuresThis indicator measures the prevention andmanagement of opthalmia neonatorum, definedas any conjunctivitis with discharge occurringduring the first two weeks of life. In manycountries where the prevalence of STIs is high andwhere eye prophylaxis is not widely practiced,ophthalmia in newborns is a common cause ofblindness. Gonococcus and chlamydia trachomitisare two leading causes of ophthalmia neonatorum.If treatment is delayed or inappropriate, theinfection may progress into systemic disease orresult in permanent eye damage (WHO, 1996a;Lawn et al., 2001).Gonococcal ophthalmitis can be prevented bycleaning the eyes immediately after birth andapplying silver nitrate solution (1%), tetracyclineNewborn Healthointment (1%), or erythromycin ointment (0.5%)within the first hour of birth. Eye prophylaxis isone of the key elements in the Essential NewbornCare Package (WHO, 1996a) and is a highly costeffectiveintervention, costing US $1.40 per caseaverted when the rate of gonococcal infection isgreater than 10% (WHO, 1991). When eyeprophylaxis fails, it is most often because it wasadministered too late (after the first hour) or theeyes were flushed after administration of silvernitrate to prevent chemical conjunctivitis, or givingdrops that were too concentrated after evaporation.How to Measure ItThis indicator is measured at the facility level in adefined geographic area and period. Datarequirements are: (1) the number of infants whoreceive eye prophylaxis within one hour of birthin a specified period; and (2) the total number oflive births in the same period. Valid data can onlybe obtained by direct observation of attendance atbirth at health facilities with high client volume.Where data on the numbers of live births for thedenominator are unavailable, evaluators canestimate the total live births from the totalpopulation and crude birth rate in a specified areaas follows:Total expected births = population x crude birth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO, 1999a, 1999b]).Strengths and LimitationsThis indicator is difficult to measure and is mostappropriate in settings where facility births are59

common. Reliable estimates for individualfacilities can only be obtained if there are largenumbers of deliveries. In developing countries,facility data are not recommended for estimatingthe proportion of babies who receive eyeprophylaxis care in the general population becausea large proportion of births occur at home; hence,facility-based data may be subject to selection bias.Surveys are not a recommended approach formeasuring this indicator as they are subject to recallbias, which is likely to increase with the length ofrecall period.60Chapter 3

OF HIV-P-PPERERCENTCENTAGEOF-POSITIVEPREGNREGNANTWOMENRECEIVINECEIVINGA COMPLETECOURSEOF ARV PROPHYLOPHYLAXISTO REDUCEMTCTIN ACCCCORDORDANANCEWITH A NATIONTIONALLALLY APPRPPROVEDTREAREATMENTPROTOCOCOLOL (ORWHO/UNAIDSSTANDANDARDSARDS) IN THE PRECEDIN12 MONTHSRECEDING 12 MUNGASS Core e Outcome IndicatorDefinitionThe percentage of HIV-positive pregnant womenreceiving a complete course of ARV prophylaxisto reduce MTCT in accordance with a nationallyapproved treatment protocol (or WHO/UNAIDSstandards) in the preceding 12 months.Numerator: Number of HIV-positive pregnantwomen receiving a complete course of ARVprophylaxis to reduce MTCT in accordancewith a nationally approved treatment protocol(or WHO/UNAIDS standards) in thepreceding 12 months.Denominator: The estimated number of HIVinfectedpregnant women giving birth in thepreceding 12 months.The definition of a “complete course of ARVprophylaxis” will depend on the country’s policyon ARV prophylaxis to reduce the risk of MTCTand may or may not include a dose for newborns.Details of the definition used should be provided.Measurement ToolsFacility-based national MIS; program monitoring;HIV sentinel surveillance-based records; NationalStatistical Office estimatesA tool for the measurement of this indicator isprovided in Monitoring the Declaration ofCommitment on HIV/AIDS: Guidelines onConstruction of Core Indicators.What It MeasuresThis is an UNGASS national program andbehavior indicator. It assesses progress inpreventing mother-to-child HIV transmissionthrough the provision of ARV prophylaxis.How to Measure ItThe number of HIV-infected pregnant womenwho have been provided with ARV prophylaxisduring the preceding 12 months in order to reducethe risk of MTCT is obtained from programmonitoring records. Only those women whocompleted the full course should be included inthe numerator. The denominator represents theestimated number of women in need of ARV, thatis, the number of HIV-infected women to whomARV prophylaxis to reduce the risk of MTCTcould potentially have been given. Thedenominator is estimated by multiplying the totalnumber of women who gave birth in the preceding12 months (central statistics office estimates ofbirths) by the most recent national estimate ofHIV prevalence in pregnant women (HIV sentinelsurveillance antenatal clinic estimates).Whether women who receive ARV prophylaxisfrom the private sector and NGO clinics shouldbe included in the calculation of the indicator isleft to the discretion of the country concerned.This decision should be based on a frank appraisalof how often ARV for pregnant HIV-infectedwomen is provided outside the government sector,and should be noted and applied consistently incalculating both the numerator and denominator.Private sector and NGO clinics that provideprescriptions for ARVs but assume that the drugswill be acquired elsewhere by the individuals arenot included in this indicator, even though suchclinics may be the major providers of services forthe reduction of MTCT. The key feature is theactual provision of the drugs.This indicator should be measured every two tothree years.Newborn Health61

Strengths and LimitationsThis indicator has the following weaknesses:(1) ANC data are often incomplete and may notreflect the true situation.(2) There may be selection bias because only thosewomen are included who self-select to accessservices.(3) Every country has its own definition of a fullcourse of ARV treatment.(4) The indicator does not assess treatmentcompliance, and, as currently defined,measures need. It does not assess whatpercentage of women accessing ANC serviceswhere PMTCT services are available actuallyavail themselves of the intervention.(5) As the number of women provided withHAART increases over time, the need forspecific ARV distribution to prevent verticaltransmission may lessen. It will be necessaryto develop specific indicators in order tocapture information on this matter.62 Chapter 3

PERERCENTOF PREGNREGNANTWOMENWHORECEIVEDATLEASEAST TWO ANTENNTENATALAL CAREVISITISITSDefinitionPercent of pregnant women who received at leasttwo antenatal care (ANC) visits.Numerator: Number of pregnant women whoreceived at least two ANC visits during aspecified period.Denominator: Total number of live births inthe same period.Measurement ToolsClinic registries or individual prenatal records(individual ANC records/cards after births);population-based survey (DHS, KPC, MICS);health facility exit interviewsWhat It MeasuresThe main purpose of this indicator is to provideinformation about women’s use of antenatal careservices. Some studies have found that women’suse of ANC is strongly associated with perinatalsurvival (McDonagh, 1996) more than it isassociated with better maternal health outcomes.ANC coverage plays an important role, therefore,in the monitoring and evaluation of programsaddressing newborn survival (Graham and Filippi,1994).How to Measure ItWhen calculating this indicator from apopulation-based survey, the numerator is thenumber of women who had a live birth in aspecified period and who reported receiving at leasttwo ANC visits during the pregnancy. Thedenominator is the total number of live births inthe same period. The number of live births in thespecified period can be a proxy for the total numberof women who needed antenatal care (ANC) inthe same period. Ideally, all births should be takeninto account when estimating this indicator. Theusual practice is to consider only live births becauseof the difficulty in obtaining information aboutnon-live births (Graham and Filippi, 1994).When survey data are used to measure thisindicator, the denominator can be the total numberof women who gave birth during the same period.Where data on the number of live births areunavailable, evaluators can calculate total estimatedlive births using census data for the totalpopulation and crude birth rate in a specified area.Total expected births = population x crudebirth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO, 1999a, 1999b]).Strengths and LimitationsThis indicator is responsive to change in the shortterm. Annual monitoring is feasible when the dataare derived from routine data sources. While thisindicator provides a crude measure of antenatalcare utilization and takes into consideration thenumber of ANC visits (Rooney, 1992), it does notcapture the timing of the visits, the reasons forseeking care, the skill of the provider, or the qualityof care received.For international comparisons, a reference periodof three to five years is probably sufficient.Evaluators should avoid frequent surveys, becausesampling error makes it difficult to assess whethersmall changes are real or due to chance variation.For comparison purpose, one must know whetherthe denominator used reflects all births (live birthsNewborn Health63

as well as non-live births), the most recent birth,or all women. Overestimation of coverage occursfrom the use of live births only in the denominator.Furthermore, observed differences in coverage maybe due not to true changes in coverage of allpregnancies, but to differences in stillbirth andabortion rates.While a birth-based analysis represents all birthsin the reference period, it over-represents womenwho have more than one birth. Women with morethan one birth are also more likely to have otherrisk factors, such as high parity and lower rates ofhealth services use. Therefore, the indicator islikely to be lower using a birth-based estimate thana women-based estimate, and this difference willbe greater the longer the reference period used.One can obtain a women-based estimate byrestricting the calculations to the most recent birth(Graham and Filippi, 1994). Because programstend to target mothers rather than births, using awomen-based denominator may be conceptuallymore appealing to program managers. However,an analysis based on all live and non-live birthsoccurring in a specified period is essential fordetermining the impact of the number of ANCvisits on pregnancy outcomes.When the indicator is calculated from routineservice statistics, the numerator may includewomen who are not counted in the denominator(i.e., those who attended two or more ANC visitsbut whose pregnancy did not result in a live birth).Routine service-based data may also suffer fromincomplete records. Information from civilregistration systems and population censuses canbe used to estimate the denominators, butpotential problems could arise if reporting isincomplete. Health facility exit interviews mayprovide a baseline measure for evaluation purposesbut are limited because they assess women whohave not completed antenatal care and who couldtheoretically still received two or more ANC visits.The content and quality of ANC visits, however,could be explored in exit interviews.64 Chapter 3

PERERCENTCENTAGEOF PREGNREGNANTWOMENWITHAT LEASTTWO DOSESOF TETETANUSTOXOIDOID VACCINEDefinitionProportion of pregnant women receiving at leasttwo doses of tetanus-toxoid (TT) vaccine duringtheir last pregnancy.Numerator: Total of TT2+TT3+TT4+TT5.Denominator: Total number of live births.Where TT2, TT3, TT4, and TT5 refer to the 2 nd ,3 rd , 4 th , and 5 th dose of tetanus-toxoid vaccineadministered (WHO, 1999a; WHO, 1999c),respectively.Measurement ToolsService statistics; population-based surveysWhat It MeasuresNeonatal tetanus is a major global public healthproblem. Despite increasing coverage of womenof childbearing age with at least two doses oftetanus toxoid in many countries, it is estimatedthat 180,000 cases of neonatal tetanus occurredin 2002, often with a high case-fatality rate(WHO, 2005). The protection of the newbornagainst neonatal tetanus is determined by theimmunization status of the mother. In order toprotect neonates, previously unimmunized womenshould receive two doses of TT or tetanusdiphtheria(Td) toxoid vaccine during their firstpregnancy and one dose of TT or Td during eachsubsequent pregnancy up to a maximum of fivedoses (Table 3.3). This measure is additional tothe use of clean practices during delivery and thecare of the infant’s umbilical cord. Protectiveantibody levels are attained in 80%–90% ofindividuals after the second dose and in 95%–98%of women after the third dose. A three-dose courseof TT or Td provides protection against maternaland neonatal tetanus for at least five years. Fifthdoses of TT or Td given later prolong the durationof immunity throughout the childbearing yearsand possibly longer.How to Measure ItFrom service statistics:The data requirements are the total number ofdoses of T2 + TT3 + TT4 + TT5 given to pregnantwomen in a reference period (usually a year) andthe number of live births in the same referenceperiod.From population-based surveys:The numerator is the number of women givingbirth during a reference period (e.g. five years) whoreport receiving at least two doses of tetanustoxoidduring their last pregnancy, and thedenominator is the number of live births in thesame reference period.The number of live births serves as a proxy for thenumber of pregnant women. Where data on thenumbers of the live births are unavailable,evaluators can estimate the total number of livebirths using census data for the total populationand crude birth rates in a specified area as follows:Total expected births = population x crude birth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO 1999a, 1999b]).Strengths and LimitationsMany national HIS routinely collect this indicatorto provide TT2+ coverage estimates for womenattending facilities for ANC or during campaigns.Newborn Health65

Most population-based surveys also collect dataon self-reported or card-documented TT coverage.Variations in the methods used to measure TT2+coverage, as well as in the definition of thenumerator and the denominator, give rise todifferences in the magnitude and reliability of theestimates obtained. For example, service statisticsrecord the total number of doses of a vaccine inthe previous 12 months, whereas survey data tendto record the total number of women who reportreceiving at least two vaccinations during the lastpregnancy in a reference period that may be up tofive years or who can show an antenatal or similarcard with TT doses recorded.Service statistics have the disadvantage that theymay be incomplete or inaccurate (WHO, 1999a).They are also subject to a selection bias and arenot representative of the general population,particularly when ANC coverage is low. However,they provide the only way of monitoring coverageon a frequent basis and may be more reliable thanself-reported data. Surveys provide the only meansto obtaining population-based coverage, butbecause many surveys rely on self-reporting, theyare subject to recall bias that is likely to increasewith the length of the recall period.Table 3.3. WHO recommended tetanus toxoid immunization schedule for women of childbearingage and pregnant women without previous exposure to TT, Td, or DTPDose of TT, Td,or DTPGivenLevel of protectionDuration ofProtectionTT 1At first contact or as early aspossible in pregnancyNoneNoneTT 2At least four weeks after TT 180%1-3 yearsTT 3At least 6 months after TT 2 orduring subsequent pregnancy95%At least 5 yearsTT 4At least one year after TT 3 orduring subsequent pregnancy99%At least 10 yearsTT 5At least one year after TT4 orduring subsequent pregnancy99%For all childbearingyears andpossibly longerTd – Tetanus-Diphtheria toxoid vaccineDTP – Diphtheria, Tetanus, Pertussis vaccine66 Chapter 3

PROPOROPORTIONOF PREGNREGNANTWOMENRECEIVINECEIVING INTERMITTENTPREVENTIVETREAREATMENTOR MALALARIAARIA PROPHYLOPHYLAXISAXIS,ACCCCORDINORDING TO NATIONTIONALPOLICYDefinitionProportion of pregnant women receivingintermittent preventive treatment or malariaprophylaxis, according to national policy.Numerator: Number of pregnant womenreceiving intermittent preventive treatment ormalaria prophylaxis, according to nationalpolicy.Denominator: Total number of pregnantwomen surveyed.Malaria medication (prophylaxis or intermittentpreventive treatment [IPT]) will vary accordingto local susceptibility and national policy. Mostcountry policies in endemic areas require that allpregnant women receive two doses of therecommended antimalarial drug (sulfadoxinepyrimethamine[SP]) at the first regularlyscheduled ANC visit after quickening (first notedmovement of the fetus) and during each regularlyscheduled visit thereafter (WHO, 2002, 2005).However, even a single dose of SP is beneficial.SP is generally more effective than chloroquinebecause of the increasing prevalence of chloroquineresistance and the need for less frequent dosingwhen compared with chloroquine. WHOpresently recommends an optimal schedule of fourANC visits, with three visits after quickening.The delivery of IPT with each scheduled visit willlikely assure that a high proportion of womenreceive at least two doses of SP.Measurement ToolsFacility records of antenatal patients; health facilitysurveys; population-based surveys; healthinformation systems (HIS)What It MeasuresThis indicator measures coverage of IPT amongpregnant women. This is one of the core indicatorsfor monitoring progress of Roll Back Malaria(RBM). Malaria is a major health risk for womenand newborn in areas where Plasmodium falciparummalaria is endemic. In stable areas of malariatransmission, malaria infection causes anemia inthe mother. The presence of malaria parasites inthe placenta also damages placental integrity andinterferes with the ability of the placenta totransport nutrients and oxygen to the fetus, therebycausing intrauterine growth retardation, a primarycause of low birth weight.Pregnant women residing in low or unstablemalaria transmission areas have a two to threefoldhigher risk of developing severe disease as aresult of malaria infection. In such areas, malariacan cause maternal death directly from infectionor indirectly by causing severe anemia. In addition,a range of adverse pregnancy outcomes, includingspontaneous abortion, still births, and congenitalmalaria, can result from malaria, causing increasedrisk of infant mortality among all babies born tomothers living in areas of unstable malariatransmission (WHO, 2002). IPT with SP duringantenatal care significantly reduces the prevalenceof maternal anemia and placental parasitemia andthe incidence of low birth weight (Steketee et al.,2001).How to Measure ItWhen data are collected by reviewing facilityrecords, or through direct observation of ANCconsultations or client exit interviews, thenumerator is the number of pregnant women givenor prescribed malaria medication in a given period.Newborn Health67

Where data on the total number of pregnantwomen are absent, WHO recommends using 3.5%of the total population as an estimate of thenumber of pregnant women (i.e., number ofpregnant women = total population x 0.035[WHO, 1999a, 1999b]).When the indicator is calculated from populationbasedsurveys, the numerator is defined as thenumber of women who were given or whopurchased malaria medication during their mostrecent pregnancy, and the denominator as thenumber of women who had a recent live birth.The time-periods for the most recent pregnancy/live birth should be specified for both thenumerator and denominator. In most surveys, thisperiod is normally restricted to three to five yearsbefore the survey.In the year 2000, the African Summit on Roll BackMalaria adopted the Abuja Declaration, whichestablished a goal that, by 2005, at least 60% ofpregnant women at risk of malaria, especially thosein their first pregnancies, should receive IPT. Thekey goal of the RBM partnership is to halvemalaria-associated mortality by 2010 and againby 2015. Target 8 of the MDGs is to have haltedby 2015 and begun to reverse the incidence ofmalaria and other major diseases. Programs/countries should be evaluated against thesebenchmarks (WHO and UNICEF, 2005).One major limitation of this indicator is thatcurrent data collection approaches lackinformation on the completeness of the drugregimen taken during pregnancy. In addition todetermining the type of malaria medication taken,information on the frequency and timing of drugadministration is required to determine whetherpregnant women are adequately protected againstmalaria. Information on the frequency and timingof drugs administered could theoretically beobtained if clinics maintain records on thenumbers of patients attending and on the numberof women given a first and second course of IPTor the number of packets of medicine disbursed.Facility records measure the proportion of womengiven or prescribed malaria medication but do notreflect the proportion of women who took themedication. Compliance with the treatment willrarely be 100% and will vary depending on manydifferent local factors. Where malaria is sporadicor seasonal, programs focus on screening womenthat present with symptoms and on treating thosewho are infected. Alternative indicators in thiscase include:• Number of pregnant women presenting withmalarial symptoms; and• Percent of pregnant women treated for malariaaccording to locally established protocols.Strengths and LimitationsSome large household surveys, such as the DHS,routinely collect data for this indicator. In additionsome health facility surveys that conduct recordreviews, direct observation of ANC consultations,or exit interviews with ANC clients yield thisinformation for client populations. The questionsasked in most population-based surveys assumethat women are able to report on malaria treatmentreliably, but few validation studies have tested thisassumption. Population-based studies also rely onself-reported data, which are subject to recall biasthat is likely to increase with the length of therecall period.68 Chapter 3

PROPOROPORTIONOF PREGNREGNANTWOMENWHOKNOW TWOOR MORENEWBEWBORNDANANGERSIGNSDefinitionProportion of pregnant women who know two ormore newborn danger signs.Numerator: Number of pregnant women whoknow two or more newborn danger signs.Denominator: Total number of pregnantwomen surveyed.“Know” refers to the ability to spontaneously namethe warning/danger signs of newborn complications.Proposed “danger signs” include:• Breathing difficulty, irregular or fast(> 60 breaths per minute)• Feeding poorly (less than half of usualconsumption)• Jaundice, pallor, bleeding• Convulsion, spasm, jitters• Fever temperature greater than 38 o C or lowtemperature less than 36 o C• Vomiting green, no stool in 24 hours of life,swollen abdomen(Lawn, McCarthy, and Ross, 2001)Measurement ToolsHealth facility survey (exit interviews);population-based surveyWhat It MeasuresThe purpose of this indicator is to assess mothers’knowledge and awareness of newborn danger signsand when to seek health care. Because most babiesare born at home or are discharged from thehospital in the first 24 hours, increasing awarenessof the danger signs of newborn complications isof critical importance for improving newbornsurvival. In many developing countries, morebabies die in the first week of life than any othertime in childhood, and those who become illshortly after birth may deteriorate and die rapidly.The warning signs of newborn illness may not berecognized, because they are often much lesspronounced that those in an older child or adult.Nevertheless, mothers and families need to knowabout danger signs of newborn illness, where togo for treatment, and the reason for respondingquickly to these danger signs.How to Measure ItThis indicator is derived from correct answersgiven spontaneously to knowledge questions askedduring a health facility exit interview or apopulation-based survey. The signs and symptomscaretakers in different settings can consistentlyrecognize should be given careful thought in eachcultural context.All pregnant women who were surveyed areincluded in the denominator, regardless of whetherthey know newborn warning/danger signs.Because the indicator is defined to reflectknowledge of a precise number of danger signs ofnewborns, individuals reporting fewer than twodanger signs are not counted in the numerator.When calculating this indicator from apopulation-based survey, the indicator is definedas the proportion of mothers who know two ormore newborn danger signs and may be restrictedto women who have had a recent live birth (forexample, the proportion of mothers with a childunder one year who know two or more newborndanger signs). The denominator is the totalnumber of mothers surveyed. The time-periodfor the most recent live birth should be specifiedNewborn Health69

for both the numerator and denominator. In mostsurveys, this period is normally restricted to threeto five years before the survey. If the sample sizeis sufficiently large, the indicator may be restrictedto women who gave birth in the year precedingthe survey.Strengths and LimitationsThis indicator is simple to measure at thepopulation level. Disaggregation of the indicatorby knowledge of individual danger signs, residence,or age group may provide useful information aboutgaps in knowledge.A major limitation of this indicator is that littleconsensus exists on which signs and symptoms thepublic can use to improve the early diagnosis ofserious illness at home. Algorithms shown to besensitive and specific in clinical settings are toocomplex for use in the general population. Simplermeasures are less specific and could lead to largernumbers of newborns receiving unnecessarytreatment. However, having some healthy babiesover-treated is preferable to having some sickbabies under-treated and dying as a result.Programs aimed at raising community awarenessof neonatal illness should carry out formativeresearch to determine what signs of illnesses arealready recognized in the community and how toadapt general recommendations to a specificsetting. More research is required to reachconsensus on which signs and symptomscaretakers in different settings can consistentlyrecognize.70 Chapter 3

PROPOROPORTIONOF DELIVERIESOCCURRINCCURRING IN A HEALEALTHFACILITCILITYDefinitionProportion of deliveries occurring in a healthfacility.Numerator: Number of deliveries occurringin a health facility in a specified period.Denominator: Total number of live births inthe same period.Measurement ToolsRoutine health service data; population-basedsurvey (DHS, KPC, MICS)What It MeasuresThe main purpose of this indicator is to provideinformation about coverage of institutionaldeliveries. Institutional delivery, especially at thetime of obstetrical emergency, and skilledattendant at birth are associated with reducedmaternal mortality (Koblinsky et al., 1999).Institutional deliveries have also been found tohave strong beneficial effects on infant survivalprobabilities (Panis, 1994).How to Measure ItWhen calculating this indicator from routineservice statistics, the numerator is the number ofpregnant women delivering at a facility in aspecified period and the denominator is theestimated total number of births in the sameperiod. The estimated number of births is a proxyfor the numbers of women who need delivery carefor a specific geographic area. Evaluators shouldcount all births but usually only use live births incalculating this indicator because of the difficultyin obtaining information about non-live births(Graham and Fillippi, 1994)Where data on the number of live births areunavailable, evaluators can calculate total estimatedlive births using census data for the totalpopulation and crude birth rate in a specified area.Total expected births = population x crude birth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO, 1999a, 1999b]).When data are derived from a population-basedsurvey, the indicator is defined as the proportionof live births delivered in a health facility in aspecified period and is based on mothers’ reportsof the place of delivery.Strengths and LimitationsMeasurement of this indicator from a populationbasedsurvey is straightforward. Annualmonitoring is only possible if data come fromroutine sources. For international comparisons, areference period of three to five years is probablysufficient. Frequent surveys are generallyundesirable because the survey periods mayoverlap, and sampling error may make it difficultto assess whether small changes are real or due tochance variation.If the indicator uses a birth-based analysis, that iscounting all births in the survey period, it will overrepresentwomen with multiple births in the surveyperiod. Women with more than one birth are alsomore likely to have other risk factors, such as highparity and lower rates of health services use.Therefore, institutional delivery coverage may beunder-estimated although the degree ofunderestimation is likely to be small.Newborn Health71

When the indicator is calculated from routineservice statistics, the numerator may includewomen not included in the denominator (i.e.,those who attended two or more ANC visits butwhose pregnancy did not result in a live birth).Routine service-based data may also suffer fromincomplete records. Information from civilregistration systems and population censuses canbe used to estimate the denominators, butpotential problems could arise if reporting isincomplete. Since the denominator for thiscalculation includes only women with live birthsand excludes women with fetal deaths and stillbirths,the only valid association will be neonatalmortality and not with perinatal mortality. It isto be noted that this indicator only measuresinstitutional delivery coverage and does notprovide any indication about the quality of care.72 Chapter 3

PROPOROPORTIONOF DELIVERIESWITH A SKILLEDATTENDTTENDANTAT BIRIRTHDefinitionProportion of deliveries with a skilled attendantat birth.Numerator: Number of deliveries with a skilledattendant at birth during a specified period.Denominator: Total number of live birthsduring the same period.Skilled attendants are individuals with “midwiferyskills (i.e., doctors, midwives, nurses) who havebeen trained to proficiency in the skills necessaryto manage normal deliveries, diagnose, andmanage or refer complicated cases” (WHO, 1999).Trained traditional birth attendants are notincluded in the definition of skilled attendant.Measurement ToolsRoutine health services data; population-basedsurveyWhat It MeasuresThis indicator measures the extent of women’s useof delivery care services. Many argue thatincreasing the proportion of deliveries with askilled attendant is the single most criticalintervention for reducing maternal mortality andincreasing newborn survival (WHO, 1999b).How to Measure ItThe data requirements are: (1) the number ofbirths attended by skilled health personnel in adefined time period; and (2) the number of livebirths in the same geographic area and referenceperiod.The number of live births is a proxy for thenumbers of women who need delivery care.Ideally, all births should be counted whencalculating this indicator. However, due to thedifficulty in obtaining information about non-livebirths, the usual practice is to only use live births(Graham and Fillippi, 1994).Where data on the number of live births areunavailable, rough approximations can be madeusing census data for the total population andcrude birth rates in a specified area as follows:Total expected births = population x crude birth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO 1999a, 1999b]).When data are derived from a population-basedsurvey, the indicator is defined as the proportionof births attended by trained medical personnelin a specified period. This is because surveyrespondents cannot assess skills.Strengths and LimitationsAnnual monitoring is feasible when data arederived from routine data sources. Forinternational comparisons, periods of three to fiveyears are probably sufficient. Frequent surveys aregenerally undesirable because the survey periodsmay overlap, and sampling error makes it difficultto assess whether small changes are real or due tochance.Differences in what definitions are used and inhow skilled attendants are reported may lead todiscrepancies between countries. Most surveyssuch as the DHS rely on women’s self-report, buthow women interpret the question on “whoassisted with delivery?” and whether theyNewborn Health73

accurately identify the health staff attending thedelivery is unknown. Even if rates of skilledattendant deliveries are similar across countries,major differences are likely to exist in howproviders are trained, in what providers are allowedto practice and do practice, and in what resources,equipment and supplies are at their disposal.As this indicator uses a birth-based analysis, thesample will over-represent women with multiplebirths in the survey period. Women with morethan one birth are also more likely to have riskfactors, such as high parity and lower rates ofhealth services use. Delivery coverage maytherefore be underestimated, although thisunderestimate is likely to be small. Furthermore,the strong correlation between skilled attendantand institutional delivery makes assessing theimpact of skilled attendant alone difficult todetermine.Evaluators can disaggregate skilled attendant atdelivery by place of delivery to further documentthe degree of care received at the time of delivery.This measure of care or “skilled attendance” willvary by setting and attendant. A skilled attendantconducting a delivery in hospital, for exampleprovides a higher level of “skilled attendance” thandoes a skilled attendant conducting a delivery athome.74 Chapter 3

MATERNTERNALMORORTALITALITY RATIOTIOTIO (MMR)DefinitionThe number of maternal deaths per 100,000 livebirths.Numerator: All maternal deaths occurringwithin a reference period (usually one year).Denominator: Total number of live birthsoccurring within the reference period.A “maternal death” (as cited in InternationalClassification of Disease [ICD]-10 [WHO,1992]) is the death of a woman while pregnant orwithin 42 days of termination of pregnancy,irrespective of the duration and the site of thepregnancy. Death can stem from any cause relatedto or aggravated by the pregnancy or itsmanagement, but not from accidental or incidentalcauses. Maternal deaths fall into two groups, directand indirect, as follows:“Direct obstetric deaths” result from obstetriccomplications of the pregnant state(pregnancy, labor, and puerperium), frominterventions, omissions, incorrect treatment,or from a chain of events resulting from anyof the above.“Indirect obstetric deaths” result from previousexisting disease or disease that developedduring pregnancy and which was not due todirect obstetric causes, but was aggravated byphysiologic effects of pregnancy.Measurement ToolsVital registration; service statistics; populationbasedsurveys; surveillanceWhat It MeasuresMaternal mortality is widely acknowledged as ageneral indicator of the overall health of apopulation, the status of women in society, andthe functioning of the health system. Highmaternal mortality ratios are thus markers ofproblems of health status, gender inequalities, andhealth service delivery in a country. The maternalmortality ratio measures obstetric risk (i.e., the riskof dying once a women is pregnant), but omitsthe risk of being pregnant (i.e., fertility in apopulation, the effect of which is reflected in thelifetime risk) (Graham and Airey, 1987). It isuseful for advocacy purposes, but lacks informationon the causes of high maternal morality or theinterventions required to reduce maternal deaths.How to Measure ItPopulation-based surveys are the primary sourceof information for calculating the maternalmortality ratio in many developing countries.These types of surveys include the following:• Reproductive Age Mortality Surveys(RAMOS) seek to identify all femaledeaths in a reproductive period, using acombination of approaches, such as crosssectionalhousehold surveys, continuouspopulation surveillance, hospital and healthcenter records, and key informants (WHO,1987).• Direct estimation, which relies on askingquestions about maternal deaths in ahousehold during a recent interval of time, sayone to two years. These questions can be askedin the context of a household survey or a censusof all households, although as yet experiencewith the latter is fairly limited (Campbell,1999).• The sisterhood method goes some way toovercoming large sample size requirements byinterviewing adult respondents about thesurvival of all their sisters. The indirectNewborn Health75

method (Graham, Brass, and Snow, 1989)involves fewer questions to respondents butprovides a pooled estimate that relatesstatistically to a point around 10-12 years priorto the survey. The direct method (Stanton,Abderrahim, and Hill, 2000) provides a morecurrent estimate at about three to four yearsprior to the survey, but requires more questionsand is more costly and time consuming.The data sources and collection methods describedabove have very different strengths and weaknessesand yield estimates of varying reliability. For thisreason, surveys to estimate maternal mortalityshould occur no more frequently than every fiveto ten years. When interpreting maternalmortality ratios, researchers must take into accountthe confidence intervals. Because of theimprecision in these estimates, modeling methodshave also been developed (WHO, UNICEF, andUNFPA, 2001; AbouZahr and Wardlaw, 2001;UNFPA, 1998). This indicator is directly relevantto Goal 5 of the MDGs, which is to improvematernal health. Target 6 set by the MDGs is toreduce the maternal mortality ratio by threequartersbetween 1990 and 2015.Strengths and LimitationsMaternal deaths are difficult to investigate becauseof their comparative rarity on a population basis,as well other context-specific factors, such asreluctance to report abortion-related deaths,problem of memory recall, or lack of medicalattribution (Campbell and Graham, 1991). Thus,no single source or data collection method isadequate for investigating all aspects of maternalmortality in all settings.Few developing countries have vital registrationsystems that are sufficiently complete to providereliable population estimates (AbouZahr, 1998).The main drawback of health services data indeveloping countries relates to the selectivity bias.Estimating the denominators for health-servicesbasedmaternal mortality rates may be challenging.Without detailed knowledge of the catchmentpopulation, it is hard to gauge whether thematernal mortality ratio underestimates oroverestimates the level for the general population.Other problems related to using health facilitystatistics include inaccuracies in routine registersand misclassification of deaths occurring outsidematernity wards. Population-based surveys canprovide up-to-date estimates but are timeconsumingand costly because they require largesample sizes to obtain single-point estimates withsufficiently narrow confidence intervals to enablemonitoring of trends. For further discussion ofthe limitations of various data sources in both thedeveloping and developed world, see AbouZahr(1999); Berg, Danel, and Mora (1996); andCampbell and Graham (1990).Due to the limitations inherent in mostmeasurement methods, maternal mortality ratiosare only a broad indication of the level of maternalmortality, rather than a precise measure. The useof confidence intervals around the estimates helpsraise awareness that a point estimate is usually tooimprecise to be used to monitor trends (Abou Zahrand Wardlaw, 2001). In addition, distinguishingbetween real and artificial changes in the maternalmortality ratio is complicated because observeddifferences do not necessarily indicate improvedmaternal health status (Graham, Fillipi, andRonsman, 1996). Other important issues toconsider include:• Non-sampling errors such as changes in theaccuracy of reporting or of classification overtime or between districts or populations(Stanton, Abserrahim, and Hill 2000);• Changes in the definition of a maternal deathbetween ICD-9 and ICD-10 (WHO, 1992,1997). Presentation of the maternal mortalityratio should thus clearly state which versionis used. In the case of ICD-10, one mustspecify which of the three categories (directand indirect maternal deaths up to 42 days76 Chapter 3

postpartum, late maternal deaths, pregnancyrelateddeaths)* are included in the numerator;• Aggregate levels may hide wide differentialsbetween population subgroups;• Apparent differences in the maternal mortalityratio between rural and urban areas may simplyreflect differences in the pattern (not level) offertility, with more rural women who are grandmultiparous and for whom the risk of deathwill likely be higher. Other possibleconfounders include general health status,such as levels of anemia or malaria, andsocioeconomic factors.* Late maternal deaths: direct or indirect obstetric causesmore than 42 days but less than one year after termination ofpregnancy. Pregnancy-related deaths: deaths while pregnantor within 42 days of the termination of pregnancy,irrespective of the cause.Newborn Health77

PROPOROPORTIONOF NEWBEWBORNSWHORECEIVETHERMALPROTECTIONIMMEDIAMMEDIATELTELY AFTERBIRIRTHDefinitionProportion of newborns who receive thermalprotection immediately after birth in a specifiedperiod.Numerator: Number of newborns receivingthermal protection immediately after birthin a specified period.Denominator: Total number of live births in thesame period.Measurement ToolsPopulation-based surveys; facility surveys;community health provider observationsWhat It MeasuresThis indicator measures the prevalence of thermalprotection of the newborn. In its current form, thisindicator is applicable in settings where mostdeliveries occur at home. Thus, the indicator is ameasure of the quality of performance of birthattendants.Thermal protection is important for full-termnewborns, but is critical for pre-term and low birthweight newborns because of increased risk of illnessand death. A newborn is most sensitive tohypothermia during the first 6-12 hours after birth.Hypothermia occurs when the body temperature ofthe newborn drops below it’s normal temperature(36.5 0 C). Hypothermia can occur if a newborn isleft wet and unprotected from cold while waitingfor the placenta to be delivered. If babies are notprotected immediately after birth, hypothermia canoccur even in a moderate or warm environmentaltemperature. As the body temperature decreases, thebaby becomes lethargic/weak, less active, hypotonic,and unable to suck. If the condition progresses, theinfant may develop serious conditions like impairedcardiac function, hemorrhage (especially pulmonary),jaundice, and eventually die.The principle ways of providing newborns withthermal protection include delivery of the baby in awarm room, drying the baby thoroughly after birth,wrapping it in a dry warm cloth while keeping it outof draughts on a warm surface, and giving it to themother as soon as possible, or by wrapping the babyin the Kangaroo care position with skin-to-skincontact with the mother. When separated from themother, a newborn baby needs to be well protectedfrom cold and/or heat. The most efficient way ofprotecting babies is to use clothes or wrap the babyin loose layers of light but warm material. If thebaby is tightly wrapped with clothes, there is littleair trapped between the body, and the cloth itselfdoes not provide sufficient insulation.How to Measure ItThis indicator needs to be measured in a definedgeographical area and period. In population-basedsurveys, the following information needs to becollected for the numerator: the number of newborns(a) who were dried and wrapped with a warm cloth,blanket, or placed in Kangaroo care, immediatelyafter birth instead of letting them rest on the flooruntil the cord was cut, or placenta delivered; (b) whowere not bathed at least for six hours after birth; and(c) whose heads were covered. In a population-basedsurvey, these data are collected from women whohad live births in a given reference period. All threecriteria ought to be met for a newborn to beconsidered as having received thermal protection.The denominator is the number of live birthsoccurring in the same reference period.Data for calculating this indicator can also becollected through direct observation of providers infacilities. Where national policy on thermalprotection of newborns exists, this should be used asNewborn Health79

a standard against which to assess the practices ofhealth care providers. If a written policy on the “warmchain” does not exist, the guidelines in Table 3.4should be used to assess the practices of providers atthe health facility level.Strengths and LimitationsThis indicator is difficult to measure and itsmeasurement in surveys is exploratory. One majorproblem is that surveys rely on recall of the events atthe time of delivery. As a result, this indicator issubject to recall bias, which is likely to increase withthe length of the recall period. Recall bias can beminimized by keeping the reference period short.This would require a bigger geographic area fromwhich to identify mothers who gave birth morerecently.Since this is a self-reported indicator, there is also apossibility that mothers would report therecommended behavior rather than actual practice.For example, in a community where the practice ofbathing the newborn is prevalent and programsaimed at raising awareness regarding newborn careexist, mothers may be aware of the correct practice,but traditional norms may prevent them fromadopting the behavior. One way to avoid this bias isthrough direct observation.In settings where births are attended by communityhealth workers or trained birth attendants, thisindicator can be measured by directly observing thehealth worker. In that case, the indicator wouldmeasure the quality of health worker performance.Although better than self-reports, direct observationis not likely to be feasible in community settings.Direct observation of health workers may only bepossible in health facilities that attend to a largenumber of births.Table 3.4. Essential newborn care intervention sub-packagePACKAGE COMPONENT-WARM CHAINAt deliveryX Ensure the delivery room is warm (25 o to 35 o C), with no drafts.X Deliver the baby on a clean surface.X Dry the baby immediately.X Wrap the baby with clean dry cloth.X Keep the baby close to the mother (ideally skin-to-skin) to stimulate early breastfeeding.X Postpone bathing for 6 hours.*After deliveryX Keep the baby clothed, wrapped with the head covered.X Minimize bathing, especially in cool weather or for small babies.X Keep the baby close to the mother.X Use kangaroo care for stable LBW babies or for rewarming stable bigger babies.X Show the family how to avoid hypothermia, how to recognize it, and how to rewarm a coldbaby. The family should aim to ensure that the baby's feet are warm.* In high HIV prevalence areas, early bathing may be a strategy to prevent MTCT/HIV.Source: Lawn, McCarthy, and Ross (2001).80 Chapter 3

TIMELIMELY INITIANITIATIONTION OF BREASREASTFEEDINTFEEDINGDefinitionProportion of infants less than 12 months of agewho were put to the breast within one hour ofdelivery (WHO, 1991).Numerator: Number of infants less than 12months of age who were put to the breastwithin one hour of delivery.Denominator: Total number of infants less than12 months of age.Measurement ToolsPopulation-based surveys employing representativesamples (e.g., DHS, MICS, KPC). Facility-basedrecords may also be used to track trends inbreastfeeding initiation among clients but not tomeasure the impact of interventions on women withinfants in the population of the catchment area.What It MeasuresThis indicator measures whether mothers in thepopulation and/or in health facilities initiate earlybreastfeeding with its respective benefits to bothmother (reduced postpartum haemorrhage) andinfant (skin-to-skin contact and exposure tomaternal antibodies in colostrum). Mothers aremore likely to successfully initiate lactation, toencounter fewer problems breastfeeding, and tomaintain optimal breastfeeding behaviors if theyinitiate breastfeeding shortly after birth.Breastfeeding should begin no later than one hourafter the delivery of the infant.have been put to the breast within one hour ofbirth.When facility data are used to calculate thisindicator, the data requirements are the numberof infants discharged from the facility during thereference period and the number of infantsdischarged who were put to the breast within onehour of birth during the same reference period. Itis important to note that the two indicators(population-level and program-level) are notcomparable.Strengths and LimitationsIn population-based surveys, mothers may havedifficulty recalling correctly when they initiatedbreastfeeding for their youngest children andwhether this was within one hour of delivery. Thisindicator may also mask changes in population orhealth facility practices that have occurred withinone year. The facility-based indicator does nothave as much recall bias, but facility-based ratescannot be used to determine population leveltrends in many settings because the data onlyreflect breastfeeding initiation by women who gavebirth in facilities.Sample Questionso Did you ever breastfeed [NAME]?o How long after birth did you first put[NAME] to the breast?How to Measure ItThe data requirements for calculating thisindicator from population-based data are thefollowing: the number of infants less than 12months of age in the population and the numberof infants less than 12 months of age reported toNewborn Health81

EXCLCLUSIVEBREASREASTFEEDINTFEEDING RATETETE (EBR)DefinitionProportion of infants aged less than 0-5 months whowere exclusively breastfed in the last 24 hours.Numerator: Number of infants aged 0-5 months(less than 180 days) who were exclusivelybreastfed in the last 24 hours.Denominator: Total number of infants aged 0-5months (less than 180 days) surveyed.Exclusive breastfeeding is the practice of only givingbreast milk to the infant, with no other solids or liquids,including water. Infants are allowed, however, to havedrops of vitamins/minerals/medicines (WHO, 1991).Measurement ToolsPopulation-based surveys employing representativesamples (e.g., the DHS, KPC) and program records ofexclusive breastfeeding rate (to track trends but notimpact)What It MeasuresThis indicator gives an overall measure of the degree towhich women have adopted behaviors consistent withthe recommendation that infants aged of 0-5 monthsshould be exclusively breastfed.* Relative to infants whoare exclusively breastfed, those who are not breastfed atall have at least 14 times the risk of death due to diarrhea.The risk is greatest in the first two months of life (Murrayet al., 1997). Even the introduction of herbal teas andwater to infants who have been exclusively breastfedincreases the risks of diarrheal morbidity and death.UNICEF and WHO recommend that all womenbreastfeed their children exclusively for the first 6 months.Newborn HealthHow to Measure ItThe data requirements are the number of living infantsunder six months of age and a 24 hour recall of all liquidsand solid food consumed by living infants less than sixmonths of age. Respondents should be probed aboutthe different types of liquids the infant may have received,including water, juice, milk, formula, and other liquids.The DHS country reports and Nutrition Reports bothpresent the exclusive breastfeeding rate (EBR) for infantsless than four months of age. However, programs cancalculate the EBR for infants less than six months ofage using DHS data.Strengths and LimitationsThis indicator should be interpreted as the percentageof infants who “are currently being exclusively breastfed,”rather than the percent that have been exclusivelybreastfed since birth. The use of a 24-hour recall periodcauses the indicator to slightly overestimate the percentof exclusively breastfed infants because some infants whoare given other liquids irregularly may not have receivedthem in the 24 hours before the survey. WHO’sIndicators for Assessing Breast-Feeding Practices, WellstartInternational’s Tool Kit for Monitoring and EvaluatingBreastfeeding Practices and Programs and the DHS reportsall calculate EBR using the 24-hour recall method.* The 2001 UN policy statement on HIV and infant feeding is asfollows: “When replacement feeding is acceptable, feasible,affordable, sustainable, and safe, avoidance of all breastfeeding byHIV-infected mothers is recommended. Otherwise, exclusivebreastfeeding is recommended during the first months of life. Tominimize HIV transmission risk, breastfeeding should bediscontinued as soon as feasible, taking into account localcircumstances, the individual woman’s situation, and the risks ofreplacement feeding (including infections other than HIV andmalnutrition). When HIV-infected mothers choose not tobreastfeed from birth or stop breastfeeding later, they should beprovided with specific guidance and support for at least the firsttwo years of the child’s life to ensure adequate replacementfeeding. Programmes should strive to improve conditions thatwill make replacement feeding safer for HIV-infected mothersand families” (WHO, 2001).83

-DPROPOROPORTIONOF WOMENWHORECEIVETWO HIGHIGH-D-DOSESUPPLEMENTUPPLEMENTS OF VITITAMINA WITHINSIXWEEKEEKS OFGIVINIVING BIRIRTHDefinitionProportion of women (both breastfeeding andnon-breastfeeding) who receive two high-dosesupplements (200,000 IU per dose) of vitamin Awithin six weeks of giving birth.Numerator: Number of women (bothbreastfeeding and non-breastfeeding) whoreceive two high-dose supplements (200,000IU per dose) of vitamin A within six weeks ofgiving birth.Denominator: Total number of women whodeliver within a given reference period.Measurement ToolsProgram statistics (usual source); population-basedsurveys, such as MICS, DHS, KPCWhat It MeasuresThis indicator is a measure of extent of protectionof the newborn against vitamin A deficiency(VAD). During lactation, maternal vitamin Arequirements rise to replace the vitamin A lostdaily in breast milk and to maintain the needs ofthe rapidly growing infants during at least the first6 months of life. Vitamin A supplementationduring lactation raises (and maintains) theconcentration of vitamin A in the breast milk ofwomen with VAD. Currently, WHO estimatesthat among children under 5 years of age, around3 million have ocular signs of VAD, and 140million have inadequate vitamin A status and areat increased risk of morbidity and mortality(WHO, 2000).Different expert groups differ on the criteria forthe “safe” infertile period after delivery duringwhich a relatively high dose of vitamin Asupplement may be given. The 1998 InternationalNewborn HealthVitamin A Consultative Group (IVACG)statement on Safe Doses of Vitamin A duringPregnancy and Lactation recommends that, inhyperendemic vitamin A-deficiency areas,breastfeeding mothers receive 200,000 IU vitaminA within eight weeks of delivery – provided thewoman is not pregnant (IVACG, 1998). Nonbreastfeedingwomen can be safely supplementedwithin six weeks of delivery. This level ofsupplementation will raise and maintain thevitamin A content of breast milk and will offsetthe depleting effect lactation may have on themother’s own vitamin A stores (ACC/SCN, 1994).To avoid confusion among health personnel aboutthe “safe infertile period,” the Pan AmericanHealth Organization (PAHO) currently advisesthat all mothers take two doses of supplement(200,000 IU per dose and at least 24 hours betweendoses) within six weeks postpartum (PAHO,2001). This recommendation is also advised byUNICEF.How to Measure ItThis indicator is usually calculated from servicestatistics, but can also be obtained for the generalpopulation from population-based surveys. Thedata requirements for calculating this indicator arethe total number of births during a given referenceperiod and the number of women receiving twohigh-dose vitamin A supplements within six weeksof delivery. Findings should be disaggregated bymother’s lactational status, and by urban/ruralresidence or socioeconomic level, if sample sizepermits.Strengths and LimitationsOne potential problem with the calculation of thisindicator is that women may deliver at a differentplace from the one where they receive thesupplementation. If the indicator is based on an85

overall figure for a district, it is generally moreaccurate than if it is based on data for specificclinics. Similarly, it is essential to specify whetherthis indicator measures supplements distributedthrough outreach workers to mothers deliveringat home or only those given at service deliverypoints.Evaluators can adapt this indicator so that it refersto all women, not just to those in the postpartumperiod, to evaluate interventions aimed at allwomen through programs such as “Healthy Days”or “National Immunization Days” (Bertrand andEscudero, 2000).An alternative indicator reflecting the adequacyof the program in meeting the needs of specificclients is the number of capsules distributed pereligible client.86Chapter 3

PRETERMBIRIRTHRATETETE (PBR)DefinitionPercentage of infants born alive before 37completed weeks of gestation per 100 live birthsin a given period.Numerator: Number of infants born alivebefore 37 completed weeks of gestation in agiven period.Denominator: Total number of live births inthe same period.“Preterm birth:” A preterm birth is defined as alive birth before 37 completed weeks of gestation.“Live birth:” A “live birth” is described by theUnited Nations (2001) as “the complete expulsionor extraction from its mother of a product ofconception, irrespective of the duration ofpregnancy, which, after such separation, breathesor shows any other evidence of life, such as beatingof the heart, pulsation of the umbilical cord, ordefinite movement of voluntary muscles, whetheror not the umbilical cord has been cut or theplacenta is attached; each product of such a birthis considered live born.”“Gestational age:” is the number of completedweeks since the last menstrual period of themother; a “full-term” baby is a baby born between39 and 42 completed weeks of gestation.Measurement ToolsPopulation-based surveys; vital registration; servicestatistics; routine HISWhat It MeasuresThis indicator is a measure of pregnancy outcome.Preterm birth has been identified as one of themost important causes of neonatal mortalityworldwide. Preterm birth accounted for about24% of neonatal deaths globally in 1999, anestimated 960,000 deaths per year (Save theChildren, 2001), and is an important determinantof neonatal morbidity, including neurodevelopmentalhandicaps, chronic respiratoryproblems, and infections. Pre-term babies mayalso find it difficult to breastfeed because thesucking reflex may not be present at birth and thestomach not sufficiently developed to accept milkimmediately.The lower the gestational age at delivery, thegreater is the chance of death or handicaps. Amoderately pre-term baby (32-37 weeks ofgestation) has a 6-20 times higher mortality ratethan a full-term baby (Kramer, 1987). A severelypreterm baby (less than 32 weeks gestation) hasup to a 100 percent chance of dying, dependingon gestation and the care available. To reducemortality in this group of babies, high-tech careis required. Mild and moderate preterm birth (32-37 weeks of gestation) is much more common andaccounts for a more significant number ofpreventable neonatal deaths than severely pretermbabies (Kramer et al., 2000).How to Measure ItTo calculate this indicator, two pieces ofinformation are needed: the number of pretermbirths in a given population and reference periodand the total number of live births in the samepopulation and reference period. The referenceperiod is usually one year but it could also be fiveyears.The preterm birth rate can be calculated at thefacility level to monitor the outcome of deliveryin health facilities using data collected by routineHIS or through record reviews (birth registers ordelivery room logs). Reliable estimates forNewborn Health87

individual facilities can only be obtained if thereare large numbers of deliveries. Facility data arenot recommended for estimating the preterm birthrate for the general population. Because a largeproportion of births occur at home in developingcountries, facility-based data may be subject toselection bias.The pre-term birth rate can also be measuredthrough a population-based survey. However, thedata may not be reliable as a valid assessment ofgestational age is often not available. Where dataon the number of live births are unavailable,evaluators can estimate the total number of livebirths using census data for the total populationand crude birth rates in a specified area as follows:Total expected births= population x crude birth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO, 1999a, 1999b]).Length of gestation is usually based on the firstday of the mother’s last menstrual period. Relyingon the date of the woman’s last menstrual periodto calculate gestational age is feasible if these dataare known and if the woman has a regularmenstrual cycle close to 28 days. Where womenoften breastfeed until their next pregnancy, the lastmenstrual period may be unknown or the woman’smenstrual cycle may be irregular, which may makethis method unreliable (Lawn et al., 2000). It mayalso be difficult to explain changes in the pretermbirth rate as trends may reflect changes in thefrequency of multiple pregnancies, obstetricintervention, increased proportion of deliveriesattended by a skilled birth attendant, increasedaccess to and use of EmOC, and changes in thenumber of deaths and increased registration/reporting of preterm babies. In addition, theavailability of consistent international comparisonsof preterm birth rates is limited.The preterm birth rate is usually calculated at thenational level. Sub-national estimates can also becalculated if sample sizes are sufficiently large.Strengths and LimitationsThis indicator is important for measuring progresstoward the prevention of preterm birth, which isa crucial strategy for improving pregnancyoutcome. However, obtaining reliable estimatesof gestational age in the general population isdifficult. Gestational age is subject to considerableerror due to recall, post-conception bleeding,irregular or long/short menstrual cycles, delayedovulation, and unrecognized fetal loss. Indeveloping countries, few babies are assessed forgestational age. Antenatal ultrasound is the “goldstandard” for assessing gestational age, but isunavailable for most women.88Chapter 3

PROPOROPORTIONOF LIVEBIRIRTHSWITH LOW BIRIRTHWEIGHTDefinitionProportion of live births with low birth weight ina specified period (e.g., 12 months).Numerator: Number of births weighing < 2500grams (g) in a specified period.Denominator: Total number of live births inthe same period.Low birth weight (LBW) is defined as a bodyweight at birth of less than 2500g.Measurement ToolsPopulation-based surveys; health services data;routine HISWhat It MeasuresThis indicator measures one of the majorobjectives of safe pregnancy/neonatalinterventions: to prevent low birth weight. LBWis also a proxy indicator to quantify the magnitudeof intrauterine growth retardation (IUGR) indeveloping countries because valid assessment ofgestational age is generally not available. IUGRis a condition in which fetal growth has beenimpaired. In developing countries, maternalunder-nutrition and maternal ill health, includingmalaria, anemia, and acute and chronic infections(e.g., sexually-transmitted infections [STIs]), aremajor causes.Low birth weight is the single most importantpredictor of newborn well-being and survival.Low-birth-weight babies are ten times more likelyto die than babies weighing over 3kg. They arealso more likely to have impaired cognitivedevelopment and to develop acute illnesses suchas diarrhea and pneumonia in early infancy (ACC/SCN, 2000). Because maternal under nutrition isa major determinant of LBW, high rates of LBWNewborn Healthshould be interpreted not only as an indicator ofnewborn under-nutrition, morbidity, andmortality, but also as an indicator of maternal wellbeing.One of the goals of the World Summit forChildren is to reduce the incidence of low birthweight to less than ten percent (ACC/SCN, 2000).How to Measure ItThe data requirements are: (1) number ofnewborns with a birth weight less than 2,500g ina defined time period (e.g. 12 months); and (2)number of live births in the same time period. Thedenominator is the number of live births occurringin the same reference period. Where data on thenumber of live births are unavailable, evaluatorscan estimate the total number of live births usingcensus data for the total population and crude birthrates in a specified area as follows:Total expected births= population x crude birth rateIn settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births = totalpopulation x 0.035 [WHO, 1999a, 1999b]).Since low birth weight is due to many complexfactors, changes in low-birth-weight incidenceoccur slowly. Estimates every five years areprobably reasonable and consistent with theschedules of many large surveys (e.g., the DHS).Evaluators must recognize that this indicator willbe slow to change, even with well-executedinterventions.Strengths and LimitationsSeveral points pertain to LBW. First, aggregatefigures of low-birth-weight incidence may hideimportant differentials between high-risk subgroups.Second, heaping of birth weight recording89

in multiples of 500g is common and affects theincidence of low birth weight. Heaping isparticularly a problem with survey data but alsoaffects facility data to some degree.* Third, surveydata rely on women’s report of their infants’ birthweight and are subject to recall bias. Validationstudies from the United States suggest thatmothers are able to recall their baby’s weightaccurately, but we are not aware of similar largescalestudies conducted in developing countries.Obtaining reliable estimates of low birth weightin the general population is difficult. In manydeveloping countries, the majority of births occurat home and babies are not weighed; thus, the datathat are available come from a relatively smallproportion of facility births. Many householdsurveys collect data on birth weight, but since theweights reported are mainly from facility births,these data are also subject to selection bias. Somehousehold surveys (such as the DHS) ask mothersto state whether their baby was smaller thanaverage or very small; and at an aggregate levelthese data may be used to estimate incidence oflow birth weight at a national level. Regionalestimates are also possible if the sample size issufficiently large (Boerma et al., 1996).* Heaping occurs when respondents do not know the exactweight. Estimated weights are often reported on certainpreferred weights, such as multiples of 100 or 500 grams.90Chapter 3

LATEFETETALDEAEATHRATETETE (LFDR)DefinitionNumber of late fetal deaths per 1,000 births (livebirths plus late fetal deaths) in a given period.Numerator: Number of babies born dead after28 weeks of gestation (or birth weight over 1kg)in a given period.Denominator: Total number of births (live birthsplus fetal deaths) in the same period.A “late fetal death” is defined as death of a fetusafter 28 weeks of gestation.The WHO definition of a “fetal death,” alsoadopted by the United Nations and the NationalCenter for Health Statistics (NCHS), is deathbefore the complete expulsion or extraction fromits mother of a product of conception, irrespectiveof the duration of pregnancy. The death isindicated by the fact that after such separation,the fetus does not breathe or show any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitivemovement of voluntary muscles.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”The terms stillbirth and fetal death are sometimesused interchangeably.Measurement ToolsPopulation-based surveys; vital registration; servicestatistics; routine HISWhat It MeasuresThe LFDR reflects directly prenatal andintrapartum care. It is a measurement of qualityof maternal health care services. There are anestimated four million late fetal deaths each year.Common causes of late fetal death are preventablematernal STIs such as syphilis, intrapartum birthasphyxia, pre-term birth, birth defects, maternalhypertension, and maternal diabetes.How to Measure ItThis indicator requires two pieces of informationfor a given population and reference period: thenumber of live-born babies and the number of fetaldeaths from 28 weeks gestation (equivalent tobirth weight over 1kg). WHO usually refers tothe expected weight at a given gestational age sincegestational assessment is often unavailable (WHO,1992). At the facility level, the numerator can bemeasured from birth registers or delivery room logsand from case reviews at the health facility (or inthe community). Some countries such as Malaysiahave nationwide systems for reporting late fetaldeaths.The denominator is the number of live births plusthe number of fetal deaths occurring in the samereference period. Where data on the number oflive births are unavailable, evaluators can estimatethe total number of live births using census datafor the total population and crude birth rates in aspecified area as follows:Total expected births= population x crude birth rateNewborn Health91

In settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO, 1999a, 1999b]).Fetal deaths or stillbirths are under-recorded inmany settings. However, they can be estimatedby using the ratio of stillbirths to early neonataldeaths, which is usually 1:1. Thus, if the numberof deaths during the first week of life is known,the number of late fetal deaths can be estimatedas this will be equal to the number of neonataldeaths. In some situations, such as high rates ofsyphilis infection, there may be more stillbirthsthan neonatal deaths.Strengths and LimitationsThis indicator suffers from under-reporting andunder recording. Health facility-based datasignificantly underestimate the problem of latefetal deaths because in many settings, many latefetal deaths and live births occur outside the healthsystem, which will cause substantial selection bias.Therefore, facility data are not recommended forestimating the LFDR for the general population.Pregnancy histories, now included in many surveysincluding the DHS, are another source of data forcalculating this indicator. However, there has beenrelatively less experience with pregnancy historiesthan with birth histories because of concerns aboutthe quality of retrospectively reported pregnancyhistories. Common problems with data qualityinclude the omission of late fetal and early neonataldeaths and difficulty in obtaining accurateinformation on gestational age or birth weight,leading to the misclassification of some stillbirthsas late spontaneous abortions.International comparisons are limited bynonstandard definitions and terminology for latefetal deaths and highly variable application of thesedefinitions. For example, ICD-10 defines late fetaldeaths from 22 to 40 weeks gestation. Forinternational comparisons, a birth weight of atleast 1000g (or of 28 weeks gestation of more ifweight is unavailable) is recommended. Errors orconfusion may arise in distinguishing between alive birth and a late fetal death.Both the coverage and quality of data on late fetaldeaths are insufficient and unreliable in manydeveloping countries due to sociocultural reasons,health system barriers, and the poor coverage andquality of vital registration systems. Socioculturalbarriers to obtaining information on pregnancyand birth outcomes include seclusion of womenand newborns at home, misconceptions about theimportance of registration and data collection, andacceptance of fetal-neonatal deaths as normal.Barriers within the health system include the lackof motivation for staff to collect necessary data,perceptions of the viability of the baby, and errorsin the coding of cause of death. Barriers toregistration include issues of accessibility andaffordability and lack of awareness of the benefitsof registration (Lawn et al., 2001).92Chapter 3

PERINERINATALAL MORORTALITALITY RATETETE (PMR)DefinitionNumber of perinatal deaths per 1000 total births(live births and fetal deaths) in a given period.Numerator: Sum of fetal deaths and deaths tolive-born babies within the first sevencompleted days (i.e., age 0-6 days) of life in agiven period.Denominator: Total number of births (livebirths and fetal deaths) in the same period.A “perinatal death” is a fetal death or an earlyneonatal death.The WHO definition of a “fetal death,” alsoadopted by the United Nations and the NationalCenter for Health Statistics (NCHS), is deathbefore the complete expulsion or extraction fromits mother of a product of conception, irrespectiveof the duration of pregnancy. The death isindicated by the fact that after such separation,the fetus does not breathe or show any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitivemovement of voluntary muscles.A “fetal death” is the death of a fetus weighing500g or more or of 22 weeks gestation or more ifweight is unavailable (ICD-10). The termsstillbirth and fetal death are sometimes usedinterchangeably. A “late fetal death” is defined asdeath of a fetus after 28 weeks of gestation.An “early neonatal death” (END) is the death of alive newborn within the first 7 completed days(i.e., 0-6 days) of life. Note: The day of birth iscounted as day 0, so that “within the first 7completed days” or “within 1 week” includes babies0-6 days old.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Great variation exists both between and withincountries on how the fetal death component ofperinatal mortality is recorded, particularly forearly fetal deaths that occur at 22 to 27-weeksgestation. For international comparison, WHO(1992) suggests including only deaths of fetusesweighing at least 1000g or of 28-weeks gestationor more if weight is unavailable. Presentations ofthe PMR should include a clear statement of thedefinition of perinatal mortality used. In practice,in most developing countries, accurate data onbirth weight or gestational age are difficult toobtain.Measurement ToolsPopulation-based surveys; vital registration; servicestatistics; routine HISWhat It MeasuresThe PMR is a key outcome indicator for newborncare and directly reflects prenatal, intrapartum, andnewborn care. It is estimated that perinatal deathsaccount for approximately 7 percent of the globalburden of disease (World Bank, 1993). The earlyneonatal component of the PMR may respondrelatively quickly to programmatic interventions,for example, following the introduction ofNewborn Health93

elements of the WHO “Essential Newborn CarePackage.” The fetal death component may declinemore slowly because it depends more oninterventions that influence primarily maternalhealth and on the availability of technologies suchas caesarian section.How to Measure ItThis indicator requires three pieces of information:the number of fetal deaths in a given populationin a given period (i.e. 12 months); the number ofdeaths of live-born babies at age 0-6 days in thesame population and period; and the number ofbirths (live births plus fetal deaths) in the samepopulation and period.The PMR obtained from large population-basedsurveys may be calculated at the sub-national levelif sample sizes are sufficiently large.Strengths and LimitationsBecause the PMR includes both fetal deaths anddeaths in the first week of life, it avoids theproblems of defining a live birth. There are,however, problems with the identification of a fetaldeath or stillbirth. According to WHO, a fetaldeath occurs after the twenty-second week ofgestation. However, different countries often useslightly different definitions, making internationalcomparisons of the PMR difficult.In many countries, vital registration data are notsufficiently complete to allow reliable estimationof the PMR. Techniques now exist for collectingdata on stillbirths, live births, and early neonataldeaths in population-based surveys throughpregnancy histories. These pregnancy historiesare now included in many surveys, including theDHS. However, there has been relatively lessexperience with pregnancy histories than withbirth histories because of concerns about thequality of retrospectively reported pregnancyhistories.Data quality is an issue. Common problems withdata quality include:94• Omission of stillbirths and early neonataldeaths. It is estimated that perinatal mortalityrates are under-reported by at least 40%(WHO and UNICEF, 1996) and by as muchas 500% in countries with high deathrates (Lumbignon, Panamonta, Laopaiboon,Pothinam, and Patithat, 1990; McCaw-Binns,Fox, Foster-Williams, Ashley, and Irons,1996).• Difficulty in obtaining accurate informationon gestational age or birth weight leading tothe misclassification of stillbirths as latespontaneous abortions.• Heaping of the reported age at death of livebirths on 7 days, leading to themisclassification of early neonatal deaths aslate neonatal deaths.*Prospective population-based surveys of pregnantwomen provide better quality data, but areexpensive to undertake.Survey-based estimates are generally subject torelatively large sampling errors, making it difficultto detect changes over short periods unless thechanges are quite large. Retrospective surveybasedestimates are often based on a five-yearperiod prior to the survey.The following caveats bear mention. The PMRis sensitive to changes in the quality of data. Forexample, a rise in the PMR may indicatedeterioration in perinatal outcomes, or animprovement in the reporting of perinatal deaths.Therefore, an assessment of data quality is anessential component of analysis. In this context,evaluators often find it useful to separate the PMRinto its two components: stillbirths and early* Heaping occurs when respondents do not know the exactage at death. Estimated ages at death are often reported oncertain preferred ages, such as seven days, leading to adistorted age distribution of deaths in which too manydeaths are reported at the preferred age, and too few at theages just before and after.Chapter 3

neonatal deaths. Data quality is generally moreproblematic for stillbirths than for early neonataldeaths, because of the ambiguity over thedefinition of fetal deaths and problems ofobtaining gestational age (WHO, 1996a).Facility data are not recommended for estimatingthe PMR for the general population because inmany settings, many perinatal deaths and livebirths occur outside the health system. Facilitybasedestimates of the PMR should also beinterpreted with caution because the rate issensitive to the types of deliveries occurring in thefacility. Consequently, the PMR may rise or fallin response to changes in the mix of deliveries in afacility. In small facilities, the PMR will bepotentially unstable because of the small numberof deliveries and perinatal deaths; thus, the PMRmay be ineffective for monitoring change over timein a single facility. Facility-based data are moreuseful for monitoring in countries where a largeproportion of births take place in facilities andwhere the completeness of routine reporting ishigh.Newborn Health95

-SCAUSEUSE-S-SPECIFICPERINERINATALAL MORORTALITALITY RATEDefinitionNumber of perinatal deaths from a specific causeper 1,000 births (live births and fetal deaths) in agiven period.Numerator: Number of perinatal deaths froma specific cause in a given period x 1,000 (or10,000 or 100,000).Denominator: Total number of births (livebirths and fetal deaths) in the same period.A “perinatal death” is a fetal death or an earlyneonatal death.The terms “still birth” and “fetal death” aresometimes used interchangeably.Specific causes that are commonly measuredinclude:• Lethal or severe congenital abnormalities;• Acute intrapartum events, resulting inintrapartum stillbirths or neonatal deaths dueto “asphyxia;” and• Infections, which may be highly specific, e.g.,syphilis infections as a cause of still births andearly neonatal deaths.The exact causes of death being coded may differdepending upon the program or locality where theindicator is being collected.The WHO definition of a “fetal death,” alsoadopted by the United Nations and the NationalCenter for Health Statistics (NCHS), is deathbefore the complete expulsion or extraction fromits mother of a product of conception, irrespectiveof the duration of pregnancy. The death isindicated by the fact that after such separation,the fetus does not breathe or show any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitivemovement of voluntary muscles.A “fetal death” is the death of a fetus weighing500g or more or of 22 weeks gestation or more ifweight is unavailable (ICD-10). The termsstillbirth and fetal death are sometimes usedinterchangeably.An “early neonatal death” (END) is the death of alive newborn within the first 7 completed days(i.e., 0-6 days) of life. Note: The day of birth iscounted as Day 0, so that “within the first 7completed days” or “within 1 week” includes babies0-6 days old.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Great variation exists both between and withincountries on how the fetal death component ofperinatal mortality is recorded, particularly forearly fetal deaths that occur at 22 to 27-weeksgestation. For international comparison, WHO(1992) suggests including only deaths of fetusesweighing at least 1000g or of 28-weeks gestationor more if weight is unavailable. Presentations ofthe PMR should include a clear statement of thedefinition of perinatal mortality used. In practice,in most developing countries, accurate data onNewborn Health97

irth weight or gestational age are difficult toobtain.Measurement ToolsFacility-based perinatal death audits; community/demographic surveillance; vital registration; verbalautopsyWhat It MeasuresThis indicator measures death from specific causesduring the perinatal period. Measurement ofcause-specific perinatal mortality is important forseveral reasons, including the following: (1) toestablish the relative public health importance ofthe different causes of death; (2) to evaluate trendsover time for specific causes of death; (3) toevaluate health interventions aimed at reducingmortality from specific causes of death; (4) toinvestigate the circumstances surrounding thedeaths in order to identify ways to reduceunnecessary death; and (5) to facilitate researchinto factors associated with mortality from specificcauses of death.How to Measure ItThis indicator can be measured at the healthfacility, district, community, or national levels,(although death coding at the community levelmay be less reliable), or a more simple surrogatemeasure of a specific cause may need to be used.Perinatal Death Audits (PNDAs) are beingincreasingly promoted in developing countries,particularly at facility level. Several trainingprograms are available for PNDAs withsupporting software, including the PerinatalEducation Program, a distance education programavailable on the internet that has been used to trainover 30,000 doctors and nurses, largely in SouthAfrica (www.pepcourse.co.za), with the PerinatalProblem Identification Program Software used toanalyze the data (www.ppip.co.za).In areas where medical certification of cause ofdeath is rare, verbal autopsy is often used toidentify the causes of death among infants andchildren. Demographic surveillance, where alldeaths are reported on a regular basis throughoutthe year may be used to identify deaths.In verbal autopsy, differing methods can be usedto get a verbal account of the cause of death. Inan open-ended history, the caregiver or next-ofkinis asked to tell about the events leading up tothe child death in their own words and probed tofollow-up on particular aspects. Close-endedquestions ask whether specific symptoms and signswere present during the final illness. “Expert”opinion or computerized algorithms are thenapplied to allocate the presumed cause of deathusing the descriptive data.Strengths and LimitationsIt is difficult to assign causes of death, even at thehealth facility level. At the community level, datacollection is often retrospective and reliant uponverbal autopsy, rather than on a clinicallydetermined cause of death. These factorscontribute to bias and may make the validity andreliability of the data questionable. Failure toenumerate all deaths can lead to an invalid measureof proportionate cause of death. For example,selective undercounting of deaths in the first hourof life will disproportionately reduce the numbersof deaths due to asphyxia and severe preterm birth.Unless the neonatal mortality rate is high, and/ora large number of births are included, the causespecificmortality rate may be misleading, as smallnumbers will not allow for the investigation of timetrends.Vital registration systems often do not havesufficient coverage to provide accurate data aboutcause-specific mortality in developing countries.Usually a 90% coverage rate is taken as a cut-offfor representation. Demographic surveillancetends to cover limited geographic areas, thus theunderlying cause-specific mortality in these areascannot be necessarily generalized to widerpopulations, as some of the populations undersurveillance are not typical due to multiple trialsand interventions.98Chapter 3

Misclassification of the cause of death not onlyaffects estimates of levels of cause-specificmortality over time, but it also compares causespecificmortality rates between two populationgroups. In mortality surveys, the accuracy of theindicator depends on the ability of respondents todescribe the final illness as well as the way in whichdiseases are understood and described in thecommunity. Clear case definitions and the use ofhierarchical categories for allocating cause of deathwill minimize subsequent errors.One limitation of cause-specific mortality rates isthat the death of a child is commonly the result ofmore than one cause. Some verbal autopsyquestionnaires, such as those developed by WHO,Johns Hopkins School of Hygiene and PublicHealth, and the London School of Hygiene andTropical Medicine, allow for multiple causes ofdeath, while others only allow for one. Wheninterpreting this indicator, it is important to knowwhether multiple causes of death are allowed forin the coding since the sum of the proportions foreach cause of death will generally be greater than1.00 when multiple causes of death are allowed.For this reason, many analyses do restrict the majorcause of death to one cause per child. The PerinatalProblem Identification Program in South Africaallows for the coding of primary obstetric causesof death (for stillbirths and neonatal deaths), finalcauses of death (in neonatal deaths), and avoidablecauses of death that are patient related, health careworker related, and administrative.Newborn Health99

BIRIRTHWEIGHTSPECIFICMORORTALITALITY RATE(B(BWSMR)DefinitionThe Birth Weight Specific Mortality Rate(BWSMR) is a stratification of a “newbornmortality rate” by birth weight grouping. Forexample, the Birth Weight Specific NeonatalMortality Rate (BWSNMR) for births weighing2500g or more is calculated as:Numerator: Number of neonatal deathsweighing 2500g or more at birth.Denominator: Total number of live birthsweighing 2500g or more at birth.And for births under 2500g, the BWSNMR iscalculated as:Numerator: Number of neonatal deathsweighing under 2500g at birth.Denominator: Total number of live birthsweighing under 2500g at birth.Measurement ToolsService statistics; HIS (in highly developedsystems)What It MeasuresBirth weight is the most sensitive predictor ofinfant survival and a good predictor of maternalhealth and well-being. The mortality rate for lowbirth weight babies is much higher than for thosewith a normal birth weight. Stratifying newborndeaths by birth weight helps to determine thecause of death and therefore to identify whereinterventions are needed. For example, deaths ofvery small babies are more likely related tomaternal causes predisposing to intrauterinegrowth retardation and preterm birth, whereasdeaths of normal birth weight babies are morelikely to be related to intrapartum asphyxia andpoor obstetric care. In the first case, interventionsshould focus on the mother (improving nutritionand reducing antenatal infection) and, in thesecond case, should focus on improving the qualityof delivery care. Evaluators can obtain additionalinformation by stratifying birth weight by time ofdeath (Table 3.5).How to Measure ItThe data requirements for calculating thisindicator are the number of deaths in a particularbirth weight grouping and the total number ofbirths in the same weight grouping.Strengths and LimitationsInformation for this indicator can only be collectedin settings where all babies are weighed. It istherefore most appropriate for use in healthfacilities but has been collected in somecommunity settings as part of maternal and perinatalhealth area surveillance systems (Lawn etal., 2001).One useful application of this type ofdisaggregation is to examine the number ofintrapartum deaths in normal birth weight babies.If the quality of obstetric care is good (and womenare not presenting very late in labor), then veryfew intrapartum deaths should occur becausedeliveries are expedited rapidly. The proportionof fetal deaths in babies of normal birth weightmay serve as a proxy indicator for intrapartumasphyxia and quality of delivery care.Newborn Health101

Table 3.5. Potential causes of death for specific age and birth weight categoriesWeightFetal DeathIntrapartumDeathEarly NeonatalDeathLate NeonatalDeathLess than2500gMaternal infection, e.g.,syphilis, other STIsComplicationsof pretermlabor/IUGRComplicationsof pretermlabor/IUGRInfection, ARIMedical complicationAPH HypertensivediseaseAsphyxiaInfectionLatecomplicationsof prematurityTetanus2500g andaboveMaternal infection, e.g.,syphilis, other STIs,malariaAsphyxia and birthtraumaAsphyxia andbirth traumaInfection, ARIMedical complicationAPH HypertensivediseaseMaternal infectionInfectionTetanus102Chapter 3

NUMBEROF NEONEONATALAL TETETANUSCASESDefinitionThe number of neonatal tetanus (NT) cases in agiven year, in a defined population, including bothsuspected and confirmed cases.A “suspected case” is any neonatal death betweenthe 3 rd and 28 th day after birth in which the causeof death is unknown; or any neonate reported ashaving suffered from neonatal tetanus between the3 rd and 28 th day after birth and not investigated.A “confirmed” case is any neonate with a normalability to suck and cry during the first 2 days oflife, and who between the 3 rd and 28 th day afterbirth cannot suck normally and becomes stiff orhas convulsions (i.e., jerking of the muscles) orboth.The basis for case classification is entirely clinicaland does not depend on laboratory confirmation.NT cases reported from hospitals are consideredconfirmed (WHO, 1999a).Measurement ToolsPopulation-based NT mortality surveys; neonataltetanus surveillance systems; population-basedsurveysWhat It MeasuresNeonatal tetanus is a major public health problemin the developing world. Neonatal tetanus isresponsible for 14% (215,000) of all neonataldeaths (WHO, 1998a). This indicator measuresachievement towards the goal of eliminatingneonatal tetanus by 2005 from the remainingcountries in which it still poses a significant diseaseburden. WHO defines elimination of tetanus asa reduction of neonatal tetanus cases to fewer thanone case per 1000 live births in every district ofevery country (WHO, 1999a).How to Measure ItThe data requirement is the number of neonataltetanus cases or deaths. In countries with tetanustoxoid (TT) immunization coverage of over 90percent and a clean delivery rate over 80 percent,the number of neonatal tetanus cases is taken asthe number of neonatal tetanus deaths reported.In countries with lower coverage, an estimate ofthe number of NT cases is based on an estimateof NT deaths calculated from the number of livebirths, the neonatal tetanus mortality rate(NTMR), TT2+ coverage and vaccine efficacy(VE) (see below).Number of NT deaths in 1 year = Live births xNTMR x (1 -TT2+ x VE)Where:NTMR = the baseline Neonatal Tetanus MortalityRate (mortality rate in unvaccinated cases)TT2+ = the proportion of pregnant womenreceiving at least two doses of TT vaccineVE = Vaccine efficacy (estimated as 0.95)The NTMR used is the latest value reported ineach country where a nationwide survey wasundertaken; if no surveys were conducted, a rateof 1, 5, 10, or 15 cases per 1000 live births isallocated on the basis of the NTMR reported incountries with similar risk factors. In LatinAmerica the WHO Regional Office (AMRO)uses a correction factor for the sensitivity of thesurveillance system to adjust for the numbers ofreported neonatal tetanus deaths (WHO, 1994a).Some countries occasionally conduct NT mortalitysurveys, and most countries with a high proportionof neonatal tetanus deaths carry out routineNewborn Health103

surveillance in “high risk” areas. Countries withNT surveillance systems assess their progressannually. However, in most cases surveillancesystems function poorly, and since high-riskpopulations for NT tend to live in rural areas withvery limited access to health care, neonatal tetanuscontinues to be seriously underreported.Community-based NT mortality surveys, forexample, suggest that routine surveillance systemsdetect only two to eight percent of all cases(WHO, 1994a). For this reason, WHOrecommends estimating the number of NT casesusing the formula presented above. Demographicsurveys, providing neonatal mortality at 4-14 dayson a 3-5 year basis, serve to evaluate surveillancedata.In countries where NT is a recognized problem,population-based surveys may provideinformation on levels and trends of neonatalmortality. These surveys provide information onneonatal mortality at 4-14 days, which is a sensitiveindicator of NT mortality (Boerma et al., 1996).Strengths and LimitationsA number of problems needs to be mentioned.First, this indicator reflects the overall magnitudeof the problem of neonatal tetanus deaths but doesnot offer a precise estimate because of seriousunderreporting from surveillance data and becauseof the many assumptions inherent in the WHOcalculation. Second, because this indicator isreported as a number rather than as a proportion,countries with lower rates of NT deaths but largerpopulations will rank ahead of countries withproportionately higher deaths rates. Third,aggregate figures at a national level may disguisepockets of high risk in certain subgroups (forexample in rural populations or low-caste groups).Surveillance systems reporting the number of NTcases should also give the percent completenessof reporting (number of NT reports received/thenumber of reports expected in the same timeperiod). Neonatal-tetanus deaths should also bereported in conjunction with TT2+ coverage andthe proportion of live births with a skilledattendant (as a proxy for proportion of cleandeliveries).104Chapter 3

NEONEONATALAL MORORTALITALITY RATETETE (NMR)DefinitionNumber of neonatal deaths per 1000 live birthsin a given period.Numerator: Number of deaths within the first28 completed days of life (0-27 days) in a givenperiod x 1000.Denominator: Total number of live births inthe same period.The NMR is often broken down into early andlate mortality rates. The Early Neonatal Mortalityrate (ENMR) is calculated as follows:Numerator: Number of deaths within the firstseven completed days of life (0-6 days) in agiven period x 1000.Denominator: Total number of live births inthe same period.The late neonatal mortality rate (LNMR) iscalculated as follows:Numerator: Number of deaths within 7-27completed days in a given period x 1000.Denominator: Total number of live births inthe same period.A “neonatal death” is defined as a death within thefirst 28 completed days of life (0-27 days).A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Note: The day of birth is counted as day 0, so that“within the first 7 completed days” or “within 1week” includes babies 0-6 days old.Measurement ToolsCensus; population-based surveys (e.g., DHS,MICS, KPC); vital registration system; servicestatisticsWhat It MeasuresThe NMR is a key outcome indicator for newborncare and directly reflects prenatal, intrapartum, andneonatal care. Early neonatal deaths are moreclosely associated with pregnancy-related factorsand maternal health, whereas late neonatal deathsare associated more with factors in the newborn’senvironment.How to Measure ItTo calculate this indicator, two pieces ofinformation are needed: the number of neonataldeaths in a given population and reference period,and the number of live births in the samepopulation and reference period. The referenceperiod is usually one year but it could also be fiveyears.Where data on the numbers of live births for thedenominator are unavailable, evaluators cancalculate total estimated live births using censusdata for the total population and crude birth ratesin a specified area.Total expected births = population x crude birth rateNewborn Health105

In a setting where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (number of live births orpregnant women = total population x 0.035[WHO 1999a; WHO 1999b]).Routine HIS may collect data for this indicatorto obtain estimates of the NMR for facilities.Facility data are not recommended for estimatingthe NMR for the general population, because inmany settings, many neonatal deaths and livebirths occur outside the health system, which willcause substantial selection bias.The NMR is usually calculated at the nationallevel. Sub-national estimates can also be calculatedif sample sizes are sufficiently large. The NMR issometimes calculated at a facility level to monitorthe outcome of delivery and newborn care inhealth facilities. Reliable estimates for individualfacilities can only be obtained for very largefacilities if there are large numbers of deliveriesand neonatal admissions.Strengths and LimitationsIn many countries, vital registration data are notsufficiently complete to allow reliable estimationof the NMR. The standard techniques forcollecting data on live births and neonatal deathsin population-based surveys have been widelyapplied in programs such as the WFS and DHS.Data quality is an important issue; commonproblems include omission of deaths, particularlyvery early neonatal deaths, and heaping of thereported age at death on 7, 28, or 30 days. *Heaping on these digits is particularly problematicbecause it will lead to the misclassification of earlyneonatal deaths as late neonatal deaths (seven days)or late neonatal deaths as post-neonatal deaths (28and 30 days).The NMR may respond fairly quickly toprogrammatic interventions, for example,immunizing all pregnant women in areas of hightetanus prevalence. However, survey-basedestimates are generally subject to relatively largesampling errors, so it is impossible to detectchanges over short periods of time unless thechanges are quite large. Also, changes in neonatalmortality rates are usually a long-termphenomenon and thus occur slowly. Therefore,we recommend collecting estimates of the NMRevery three to five years or longer.One limitation is that the NMR is sensitive tochanges in data quality. For example, a rise in theNMR may indicate deterioration in newbornhealth outcomes, or it may indicate animprovement in the reporting of neonatal deaths.Therefore, assessing data quality is essential toanalysis.Also, comparisons of facility-based estimates ofthe NMR should be interpreted carefully becausethe NMR in a facility is very sensitive to the casemix of deliveries and neonatal admissions. Ahigher NMR in one facility may not suggestpoorer quality of neonatal care in that facilitybecause the NMR may rise or fall with changes inthe case-mix. Also, improvements in prenatal andintrapartum care and advances in medicaltechnology may increase the NMR because babieswho may otherwise have been stillbirths maysurvive delivery only to die in the neonatal period.For these reasons, we recommend that evaluatorsbreak down facility-based estimates of the NMRby birth weight (see Birth Weight SpecificMortality Rate) and by admission status (directadmission or transfer-in) as a proxy for case mix.*Heaping occurs when respondents do not know the exact age atdeath. Estimated ages at death are often reported on certainpreferred ages, such as 7, 28, or 30 days, leading to a distorted agedistribution of deaths in which too many deaths are reported atthese preferred ages, and too few at the ages just before and after.106Chapter 3

RefererencesencesAbouZahr C. (1998). Maternal Mortality Overview. In: Murray CJL and Lopez AD (eds). HealthDimensions of Sex and Reproduction. Global Burden of Disease and Injury Series III, 111-164. Geneva,Switzerland: World Health Organization.AbouZahr C. (1999). Measuring Maternal Mortality: What Do We Need To Know? In: Berer M andRavindran TKS (eds). Safe Motherhood Initiatives: Critical Issues: Reproductive Health Matters, 13-23. Oxford:Blackwell Science Ltd.AbouZahr, C. and Wardlaw T. (2001). Maternal Mortality at the End of the Decade: Signs of Progress?Geneva, Switzerland: World Health Organization. Bulletin of the World Health Organization.ACC/SCN News. (1994). Maternal and Child Nutrition. Number 11.ACC/SN. (2000). Pojda J, and Kelly L (eds.). Low Birthweight: Report of a Meeting in Dhaka, Bangladesh14-17 th June 1999. Geneva, Switzerland: ACC/SCN in collaboration with ICDDR, B. Nutrition PolicyPaper, No. 18.Bang AT, Bang RA, Baitule SB, Reddy MH, and Deshmukh MD. (1999). Effect of Home-based NeonatalCare and Management of Sepsis on Neonatal Mortality: Field Trial in Rural India (comments). Lancet,354, 9194:1955-1961.Berg C, Danel I, and Mora G. (1996). Guidelines for Maternal Mortality Epidemiological Surveillance.Washington, D.C.: OPS.Bertrand JT, and Escudero G. (2003). Compendium of Indicators for Evaluating Reproductive Health Programs.Chapel Hill, NC: MEASURE Evaluation.Bertrand JT, and Tsui A (eds.). (1995). Indicators for Reproductive Health Program Evaluation. Chapel Hill,NC, University of North Carolina at Chapel Hill: Carolina Population Center, EVALUATION Project.Boerma JT, Weinstein KI, Rutstein SO and Sommerfelt AE. (1996). Data on Birthweight in DevelopingCountries: Can Surveys Help? Geneva, Switzerland: World Health Organization. Bulletin of the WorldHealth Organization, 74, 2:209-216.Campbell O. (1999). Measuring Progress in Safe Motherhood Programmes: Uses and Limitations ofHealth Outcome Indicators. In: Berer M and Ravindran TKS (eds.). Safe Motherhood Initiatives: CriticalIssues: Reproductive Health Matters, 31-42. Oxford: Blackwell Science Ltd.Campbell OMR, and Graham WJ. (1990). Measuring Maternal Mortality and Morbidity: Level and Trends.London, England: Maternal and Child Epidemiology Unit. Publication No. 2. London School of Hygieneand Tropical Medicine.Campbell OMR, and Graham WJ. (1991). Measuring Determinants of Maternal Mortality and Morbidity:Defining and Selecting Outcomes and Determinants, and Demonstrating Association. London, England: Maternaland Child Epidemiology Unit, London School of Hygiene and Tropical Medicine. Publication No 4.Newborn Health107

Graham W, and Airey P. (1987). Measuring Maternal Mortality: Sense and Sensitivity. Health Policy andPlanning, 2(4):323-333.Donnay F. (2000). Maternal Survival in Developing Countries: What Has Been Done, What Can BeAchieved in the Next Decade. International Journal of Gynaecology and Obstetrics, 70(1):89-97.De Brouwere V, Tonglet R, and Van Lerberghe W. (1998). Strategies for Reducing Maternal Mortality inDeveloping Countries: What Can We Learn from the History of the Industrialized West? Tropical Medicineand International Health, 3:771-782.Espeut D. (1998). Perinatal Mortality in Developing Countries: A Review of the Current Literature andMethodological Issues in Community-based Assessment. Arlington, VA: John Snow Inc. MotherCare TechnicalWorking Paper No. 5.Gloyd S, Chai S, and Mercer MA. (2001). Antenatal Syphilis in Sub-Saharan Africa: Missed Opportunitiesfor Mortality Reduction. Health Policy and Planning, 16(1):29-34.Goodburn EA. (2002). Using Process Indicators to Monitor and Evaluate Obstetric Services in DevelopingCountries. Journal of the American Medical Women’s Association, 57(3):145-8.Graham W, and Filippi V. (1994). Monitoring Maternal Health Goals: How Well do the Indicators Perform?London, England: Maternal and Child Epidemiology Unit, London School of Hygiene and TropicalMedicine.Graham W, Brass W, and Snow RW. (1989). Estimating Maternal Mortality: The Sisterhood Method.Studies in Family Planning, 20(3):125-135.Graham W, Fillipi V, and Ronsmans C. (1996). Demonstrating Programme Impact Using MaternalMortality. Health Policy and Planning, 11(1):16-20.IVACG. (1997). IVACG Statement: Maternal Night Blindness: Extent and Associated Risk Factors. Washington,DC: International Vitamin A Consultative Group. Retrieved from http://www.who.int/vaccines-diseases/en/vitamina/PDF/13_MATERNAL_NIGHT.PDFIVACG. (1998). IVACG Statement: Safe Doses of Vitamin A during Pregnancy and Lactation. Washington,DC: International Vitamin A Consultative Group. Retrieved from http://ivacg.ilsi.org/file/a5_pregnancy.pdfKoblinsky MA, Campbell O, and Heichelheim J. (1999). Organizing Delivery Care: What Works for SafeMotherhood? Bulletin of the World Health Organization, 77(5):399-406.Kramaer MS. (1987). Determinants of LBW: Methodological Assessment and Meta-Analysis. Bulletin ofthe World Health Organization, 65(5):663-737.Kramer MS, Demissie K, Yang H, Platt RW, Sauve R, and Liston R. (2000). The Contribution of Mildand Moderate Preterm Birth to Infant Mortality. Fetal and Infant Health Study Group of the CanadianPerinatal Surveillance System. Journal of the American Medical Association, 284(7):843-849.108Chapter 3

Lawn J, McCarthy B, and Ross SR. (2001). The Healthy Newborn: A Reference Manual for Program Managers.CARE, Centers for Disease Control. CCHI and World Health Organization.Lumbignon P, Panamonta M, Laopaiboon M, Pothinam S, and Patithat N. (1990). Why Are Thai OfficialPerinatal and Infant Mortality Rates So Low? International Journal of Epidemiology, 19:997-1000.McCaw-Binns AM, Fox K, Foster-Williams KE, Ashley DE, and Irons B. (1996). Registration of Births,Stillbirths and Infant Deaths in Jamaica. International Journal of Epidemiology, 25:807-813.McDonagh M. (1996). Is Antenatal Care Effective in Reducing Maternal Morbidity and Mortality?Health Policy Planning 11,1:1-15MEASURE DHS+. (2000). Service Provision Assessment (SPA) Tool Kit. Baltimore, MD: Macro InternationalInc.MEASURE DHS+. (2001). Service Provision Assessment Guidelines and Field Instruments (Unpublished).Calverton, Maryland: ORC Macro.Ministry of Health [Kenya], National Council for Population and Development [Kenya], and ORC Macro.(2000). Kenya Service Provision Assessment Survey 1999. Calverton, Maryland: Ministry of Health, NationalCouncil for Population and Development, and ORC Macro.O’Gara C, Newsome MH, and Viadro C. (1995). Indicators for Reproductive Health Evaluation: FinalReport of the Subcommittee on Breastfeeding. Chapel Hill, NC: Carolina Population Center.PAHO. (2001). Integrated Vision for Vitamin A Supplementation in the Americas. Washington, D.C.: PanAmerican Health Organization. Regional Meeting Report.Panis CW, and Lillard LA. (1994). Health Inputs and Child Mortality: Malaysia. Journal of HealthEconomics, 13(4):455-89.Safe Motherhood Inter-Agency Group. (1997). The Safe Motherhood Action Agenda: Priorities for the NextDecade. Report on the Safe Motherhood technical consultation. October 1997, Colombo, Sri Lanka.Retrieved from http://www.safemotherhood.org/resources/pdf/e_action_agenda.pdfSave the Children. (2001). The State of the World’s Newborns. Washington, D.C.: Save the Children Fund.Stanton C, Abderrahim N, and Hill K. (2000). An Assessment of DHS Maternal Mortality Indicators.Studies in Family Planning, 31(2):111-123.Steketee RW, Nahlen BL, Parise ME, and Menendez C. (2001). The Burden of Malaria in Pregnancy inMalaria-endemic Areas. American Journal of Tropical Medicine and Hygiene, 64,1/2:S28-35.Saugstad OD, Rootwelt T, and Aalen O. (1998). Resuscitation of Asphyxiated Newborn Infants withRoom Air or Oxygen: An International Controlled Trial: The Resair 2 Study. Pediatrics, 102z(1):E1.Newborn Health109

United Nations. (2001). Principles and Recommendations for a Vital Statistics System Revision 2. New York:United Nations. Sales No. E.01.XV11.10.UNAIDS. (2002). Monitoring the Declaration of Commitment on HIV/AIDS: Guideline on Construction ofCore Indicators. Geneva, Switzerland: UNAIDS. Retrieved from http://www.unaids.org/UNGASS/docs/JC718-CoreIndic_en.pdfUNFPA. (1998a). Indicators for Population and Reproductive Health Programmes. New York, NY: UNFPA,Technical and Policy Division.UNICEF/WHO/UNFPA. (1997). Guidelines for Monitoring and Evaluating the Availability and Use ofObstetric Services. New York: United Nations Children’s Fund.Villar J, and Khan-Neelofur D. (2000). Pattern of Routine Antenatal Care for Low-risk Pregnancy. CochraneDatabase Systems Review, 2:CD000934.WHO. (1987). Studying Maternal Mortality in Developing Countries: A Guidebook. Geneva, Switzerland:World Health Organization. WHO/ FHE/87.7.WHO. (1991). Indicators for Assessing Breastfeeding Practices. Geneva, Switzerland: World HealthOrganization. Report of an Informal Meeting. Division of Diarrhoeal and Acute Respiratory DiseaseControl, June 11-12.WHO. (1991b). Maternal and Perinatal Infections: A Practical Guide. Geneva, Switzerland: World HealthOrganization. Report of a WHO Consultation. Maternal Health and Safe Motherhood Programme,Division of Family Health.WHO. (1992). International Statistical Classification of Disease and Related Health Problems, 10 th Revision.Geneva, Switzerland: World Health Organization. WHO 1:1238.WHO. (1996a). Mother-Baby Package: Implementing Safe Motherhood in Countries. Geneva, Switzerland:World Health Organization. Maternal Health and Safe Motherhood Program, Division of Family Health.WHO/FHE/MSM/94.11.WHO. (1996c). Indicators for Assessing Vitamin A Deficiency and Their Application in Monitoring and EvaluatingIntervention Programs. Geneva, Switzerland: World Health Organization. WHO Nutrition 96.10.WHO. (1997). Monitoring Reproductive Health: Selecting a Short List of National and Global Indicators.Geneva, Switzerland: World Health Organization, Division of Family and Reproductive Health.WHO. (1998a). Postpartum Care for the Mother and Newborn: A Practical Guide. Geneva, Switzerland:World Health Organization, Maternal and Newborn Health/Safe Motherhood Unit, Division ofReproductive Health (Technical Support). WHO/RHT/MSM/98.3.WHO. (1998b). Basic Newborn Resuscitation: A Practical Guide. Geneva, Switzerland: World HealthOrganization. WHO/RHT/MSM/98.1. Retrieved from http://www.who.int/reproductive health/publications/MSM_98_1_table of contents.en.html110Chapter 3

WHO. (1999). A Joint WHO/UNFPA/UNICEF/World Bank Statement on Reduction of Maternal Mortality.Geneva, Switzerland: World Health Organization.WHO. (1999a). Indicators to Monitor Maternal Health Goals: Report of a Technical Working Group, Geneva,8-12 November 1993. Geneva, Switzerland: World Health Organization, Division of Family Health.WHO. (1999b). Reduction of Maternal Mortality: A Joint WHO, UNFPA, UNICEF, World Bank Statement.Geneva, Switzerland: World Health Organization.WHO. (1999c). Field Manual for Neonatal Tetanus Elimination. Geneva, Switzerland: World HealthOrganization, Department of Vaccines and Other Biologicals. WHO/V&B/99.14. Retrieved from http://www.who.int/vaccines-documents/DocsPDF/www9563.pdfWHO. (2000). Nutrition for Health and Development. Geneva, Switzerland: World Health Organization.WHO/NHD/006.WHO (2000a). Maternal and Neonatal Tetanus Elimination by 2005: Strategies for Achieving and MaintainingElimination. Geneva, Switzerland: World Health Organization. WHO/V&B/02.09. Retrieved fromhttp://www.who.int/vaccines-documents/DocsPDF02/www692.pdfWHO. (2001). New Data on the Prevention of Mother-to-Child Transmission of HIV and Their PolicyImplications: Conclusions and Recommendations World Health Organization: Geneva, 2001. TechnicalConsultation on Behalf of the UNFPA/UNICEF/WHO/UNAIDS Inter-Agency Task Team on Motherto-ChildTransmission of HIV. WHO/RHR/01.28.WHO. (2001a). Fifty-Fourth Word Health Assembly, May 14-22, 2001. Geneva, Switzerland: World HealthOrganization.WHO. (2002). Strategic Framework for Malaria Control during Pregnancy in the WHO Africa Region. [FinalDraft.]WHO. (2005). Global Immunization Vision and Strategy: Facts and Figures. Geneva, Switzerland:World Health Organization. Retrieved from http://www.who.int/vaccines/GIVS/english/Global_imm._data_EN.pdfWHO and UNICEF. (1996). Revised 1990 Estimates of Maternal Mortality: A New Approach by WHO andUNICEF. Geneva, Switzerland: World Health Organization.WHO. (2000). Framework for Monitoring Progress and Evaluating Outcomes and Impact. Geneva, Switzerland:World Health Organization. WHO/CDS/RBM/2000.25. Retrieved from http://whqlibdoc.who.int/hq/2000/WHO_CDS_RBM_2000.25.pdfWHO, UNICEF, and UNFPA. (2001). Maternal Mortality in 1995: Estimates Developed by WHO, UNICEFand UNFPA. Geneva, Switzerland: World Health Organization. WHO/RHR/01.9.World Bank. (1993). 1993 World Development Report: Investing in Health. New York, NY: Oxford UniversityPress.Chapter 3. Newborn Health111

4IMMUNIZMMUNIZATIONIONIndicators:• Proportion of infants born protected against neonatal tetanus• BCG coverage• DTP1 coverage• DTP3 coverage• OPV3 coverage• Measles coverage• HepB3 coverage• Hib3 coverage• Dropout from DTP1 to DTP3• Fully immunized child (FIC)• Vaccine wastage rate

HAPTER 4. ICHAPTER4. IMMUNIZAMMUNIZATIONChild immunization is one of the most costeffectivepublic health interventions forreducing child morbidity and mortality. Theultimate goal of immunization programs is toreduce the incidence of vaccine-preventablediseases in children by attaining high levels ofcoverage with potent vaccines administered at theappropriate ages (and recommended intervalsbetween doses for multiple dose vaccines). Thetraditional six target diseases are poliomyelitis,diphtheria, tetanus, pertussis, measles, andtuberculosis. WHO has recommended that thehepatitis B vaccine be included in nationalschedules for delivering the primary series ofvaccines to children under one year. In manyAfrican countries, national immunization policyincludes giving yellow fever vaccine at the sametime as measles. Countries are also beingencouraged to introduce new and previouslyunderutilized vaccines such as HaemophilusInfluenzae Type B. These vaccinations form partof the basic childhood immunization package incountries where they are appropriate or whereresources are available.Immunization program managers and serviceproviders need continuous information to answerthe following questions (USAID, 2003):(1) Are immunization services accessible to thetarget population?(2) How many individuals in the target populationare being vaccinated? Who is not beingvaccinated and why?(3) Does the quality of services meet programstandards?(4) Are resources being used efficiently?(5) Are service strategies meeting their objectives?(6) Are mortality and morbidity from routinediseases being reduced?Table 4.1 provides an illustrative framework formonitoring immunization programs. Monitoringin this context is keeping a close watch on thevarious functional or operational aspects ofimmunization programs related to routineimmunization, new vaccine introduction, polioeradication, surveillance, and so forth. The criticalinputs for monitoring immunization programs arevaccines, refrigerators, temperature charts, needles,syringes, and so forth. Outputs may includeimmunization sessions, health education sessions,outreach, and quality services. The main outcomeis immunization coverage. The intended impactis reduced incidence of vaccine-preventablediseases and lower infant and child mortality.Evaluations are carried out periodically to answerthe following kinds of questions (USAID, 2003):(1) What do clients, health workers, managers,and/or other stakeholders think about theservice or specific aspects of it? What do theylike? What do they dislike?(2) Were stated objectives achieved? How?(3) Which inputs led to improvements and whichdid not? Were there any unintended, positiveoutcomes? What were they? How can they bereplicated?(4) How efficiently were activities implemented?(5) Which strategies should be continued?An evaluation can undertake a comprehensivestudy of the whole health system, or focus on aservice or program within the system, or a singlefunction such as cold chain, disease surveillance,or training. Health service evaluations usuallyfocus on outcomes (e.g. changes in immunizationcoverage), processes (e.g. vaccine delivery), and/or client satisfaction. Impact evaluations examinethe effects of activities on morbidity and mortalitybut are less common because they are timeChapter 4. Immunization113

Table 4.1. Illustrative inputs, processes, outputs, and outcomes for monitoring immunization programsInputsProcessesOutputsOutcomesImpact• Vaccines• Refrigerators• Temperaturecharts• Vaccinationcards• Needles,syringes• Etc.• Training• Supervision• Servicedelivery• Surveillance• Etc.Functional Outputs• Immunizationsessions held• Educationsessions held• Health workerstrained in EPI• Etc.Service Outputs• Quality services• Clientsatisfaction• Increasedcoverage• Reduceddropout• Increasedparents'knowledgeof when toreturn• Reduceddiseaseincidence• Lower infantand childmortalityconsuming and costly and because of the difficultyof determining cause and effect relationships.Measurement Tools and Data SourcesMany tools are available for identifyingimmunization service delivery problems and howbest to address them. Routine monitoring toolsinclude patient registers, vaccination cards, ticklerfiles, tally sheets, and immunization monitoringcharts. Patient registers are used to identifychildren who are due for a vaccination, monitormissed opportunities, check the accuracy ofreporting and target case investigations.Vaccination cards and home-based records enablecaretakers and health workers to monitor a child’sprogress towards full immunization. These toolsare valuable if patient records are poorlymaintained or if a child moves from one facilityto another. Tickler files are boxes in whichchildren’s vaccination cards are filed according tothe month in which the next vaccination is dueand aid in monitoring missed opportunities forvaccination. Tally sheets are forms on which healthworkers make a mark every time a vaccination isadministered and which are used for reporting tothe district level and monitoring the accuracy ofreporting from health facilities to the district.Immunization charts monitor a health facility’sprogress toward coverage objectives.Other methods that can be used to monitor andevaluate program performance includeimmunization program reviews (conductednationally following guidelines available fromWHO); system reviews of the cold-chain orvaccine logistics systems (methods available fromWHO); health facility surveys (for informationconcerning the availability of vaccines, essentialequipment and supplies, and observations ofhealth worker vaccination practices); and reviewsof vaccine safety (methods available from WHO).System reviews of the cold-chain and vaccinelogistics system look at refrigerator temperature,storage facility adequacy, and injection safety.Health facility surveys and periodic reviews ofimmunization practices use direct observation to114Chapter 4

examine counseling, vaccine safety, availability ofessential vaccines, equipment, and supplies.Supervisory-based monitoring has also been usedin many countries to monitor and evaluate healthworker practices. Tools such as the ImmunizationServices Assessment Guide (sometimes called “thecommon assessment tool”) are available fromWHO/EPI for adaptation and enable users toconduct a comprehensive assessment of all levelsof immunization services and to examine theirrelationship to the health system as a whole(USAID, 2003).Sources of data for calculating immunizationcoverage are routine reports (also referred to as“program statistics” or “administrative data”) andhousehold surveys. Small sample surveys usinglot quality assurance methods have been used insome places to assess the quality of service, trackcoverage over time, and validate immunizationcoverage at individual health service units, or both,for the purpose of directing attention and supportto the facilities or areas that need it most. It isoften used in areas that do not correspond toofficial reporting sites, such as urban slums(USAID, 2003). Periodic evaluation of populationcoverage is frequently conducted using large-scalepopulation-based surveys such as the DHS, theKPC, the Arab League’s PAPCHILD, CDC’sReproductive Health Survey, or UNICEF’sMultiple Indicator Surveys. Many of these surveysuse either basic cluster designs or complexstratified sampling designs.Other population-based surveys includeimmunization coverage surveys, seventy-fivehousehold surveys, and missed opportunitysurveys. Immunization coverage surveys use astandard WHO methodology for determiningimmunization coverage based on a survey of asmall number of individuals (210 in 30 clusters ofseven individuals each). Seventy-five householdsurveys focus on households that have easy accessto health facilities. These surveys are based onthe assumption that if people in the 75 to 100households that are closest to the health facilitiesare not receiving services, then use of services inthe wider catchment area must be poor. Seventyfivehousehold surveys are useful in areas wherethe population is stable and coverage is unknown(USAID, 2003; WHO Training for Mid-levelManagers: Increase Immunization Coverage, AnnexC). While surveys are important they are not goodtools for routine monitoring as they are conductedperiodically (every three to five years). Therefore,immunization programs rely on routine data tomeasure coverage.Surveillance of vaccine-preventable diseases (andadverse events following immunization) is animportant component of immunization programsand can be used as a tool to identify the presenceof vaccine-preventable diseases and guide actionsto prevent them from becoming public healthproblems. Many countries have had verticaldisease surveillance programs for polio andtuberculosis. Recognizing that a fragmentedapproach to disease surveillance can be costly,inefficient, and result in duplication of effort, manycountries are now moving towards integratedsystems of disease surveillance and response.WHO-AFRO and CDC have developed a set ofindicators for monitoring and evaluating thequality of integrated disease surveillance andresponse activities, which are discussed in the nextchapter.Methodological Challenges of EstimatingCoverageFollowing are some key methodological challengesof estimating immunization coverage.Immunization coverage rates are usually based onroutine data derived from tally sheets that are filledout at the health facility level. Although surveysare not the primary tool for monitoringimmunization programs, both routine and surveybaseddata are covered in this section. Coveragerates can vary greatly by source of data. Users haveto be aware, therefore, of the strengths andlimitations of each data source in order to makesense of any data.Immunization115

Routine data are relatively inexpensive but have anumber of weaknesses that may result in invalid orunreliable estimates of immunization coverage.Routine data are based on data collected by healthfacilities and other providers on the number ofchildren immunized with specific vaccines. Thesedata are available monthly and can be used at everyadministrative level. The data are usuallyaggregated at the district level, which is preciselythe level at which managers have responsibilityfor improving the performance of immunizationservices. Routine data, therefore, are moreappropriate than surveys for active monitoring ofimmunization programs. In addition, routine dataare relatively inexpensive and can add additionalelements for little marginal cost, as opposed tosurveys.However, routine data have a number ofweaknesses that concern both the numerator anddenominator. At the facility and district levels,pressures to achieve targets may result in anupward bias in the reporting, while a lack ofinterest in record keeping and reporting may leadto underestimates of coverage (Bos and Batson,2000). These errors are compounded when districtor regional estimates are aggregated at the centrallevel.Another weakness lies in the estimation of thedenominators for routine-based coverage rates.Complete vital registration is the most reliablesource of data for the estimation of thedenominators for routine-based coverage.However, vital registration systems are incompleteand of poor quality in many low- and middleincomecountries. Estimates of the denominatorsare therefore based on counts or estimates by localhealth personnel, or on projections from the latestcensus data. Use of the latest census data canintroduce considerable uncertainty depending onhow long ago the census was conducted.Projections are usually made with cohortcomponentmethods for which estimates offertility and mortality rates are required. Censusestimates of the number of women of reproductiveages are then multiplied by the age-specific fertility116rates to obtain the number of births. Estimates ofinfant mortality are then used to reduce thenumber of births to obtain an estimate of thesurviving number of children 0-11 months of age(Bos and Batson, 2000).At the district level, estimates of the denominatorsmay also be affected by migration. As a result ofnet in-migration, districts may report routinecoverage rates greater than 100 percent of theassumed target population. The difficulty ofproviding accurate denominators is a majorobstacle to obtaining accurate nationalimmunization coverage estimates from routinedata (Bos and Batson, 2000).The quality of routine immunization data may beaffected by problems related to the accuracy,completeness, and timeliness of reporting.The reliability of routine immunization coverageestimates depends in large measure on the quality,accuracy, timeliness and completeness ofadministrative immunization reporting systems.The lack of completeness of reporting may alsohinder the compilation of reports for givenreporting periods. The WHO document “EPIInformation System Global Summary, September1998” included an article that examined theconsistency and reliability of reported routine childimmunization coverage estimates from 217countries and territories worldwide over the period1991 to 1996. That study found that 24 percentof the expected 6,000 reports were missing. TheImmunization Data Quality Audit to beperformed in the context of the Global Alliancefor Vaccines and Immunization Monitoring andEvaluation Activities (WHO, 2002) provides amechanism for determining the reliability of acountry’s routine immunization reporting system.Sampling errors from survey-based estimates tend tobecome larger at the sub-national level making itdifficult to compare different districts or measurechanges over time.A survey is geographically representative only atthe level at which the sample is drawn. Due toChapter 4

cost issues (the larger the sample size, the moreexpensive the survey), it is often not possible todraw samples that are representative at the districtlevel. Sampling errors (used to constructconfidence intervals to indicate a probability thatthe true estimate of coverage is within theseintervals) tend to be larger at the district level,making it difficult to compare different districtsor measure district-level changes in coverage overtime. Although they are often available fromsurveys, regional estimates may mask district-leveldisparities, especially when coverage is low.Therefore, the utility of nationally representativesurveys for monitoring immunization coveragemay be low at the district level where EPI mangershave the responsibility to take appropriate actionsto improve service delivery.Low availability of health cards makes it difficult touse survey-based data to estimate the timeliness ofimmunization coverage.Survey findings on immunization rely on bothhealth cards presented by mothers and history tomeasure whether and when vaccines were received.If filled out correctly, health cards are reliablerecords of children’s immunization coverage or ofmothers’ receipt of tetanus toxoid immunization.Because not all mothers can produce a health cardat the time of the survey, they are also asked abouttheir receipt of tetanus toxoid injections or thechild’s receipt of recommended vaccines. Whilemother’s recall of a child’s immunization historyhas been found to be quite accurate in a numberof studies, mothers may not remember accuratelywhich vaccines or how many doses (of a multipledosevaccine) their children have received, nor theexact dates of vaccination (Goldman and Pebley,1994). Poor recall may result in biased estimatesof immunization coverage and make it difficult toestimate accurately the timeliness ofimmunization.The timing of surveys relative to NationalImmunization Days (NIDs) for polio immunizationor large-scale measles immunization campaigns mayaffect coverage rates.If NIDs or measles campaigns occur just beforeor during the period a survey is conducted in thefield, or just after the survey has been completed,coverage rates may be affected. In the first case,coverage rates may be biased upwards, especiallyif efforts are not sustained. If a large immunizationcampaign is conducted right after the completionof the survey, there may be a discrepancy betweenthe survey- and the campaign-based estimates.Using DTP3 coverage rates avoids this potentialissue, as DTP is generally not included incampaigns.Survey data are associated with an inbuilt time lagin coverage estimates.Although population-based surveys can be usedfor estimating the proportion of the populationprotected against vaccine-preventable diseases,they are collected every three to five years. Surveystypically provide coverage estimates for children12-23 months of age and include the vaccinationsthat these children receive up to the age of 12months. This is the recommended methodologyfor survey-based estimates for immunizationcoverage. Some surveys include the vaccinationsthat children receive up to the time of the survey,which means that vaccinations given to childrenwhen they were older than 12 months of age areincluded, thereby inflating immunizationcoverage. In both cases, survey estimates provideinformation about immunization coverage at least12-23 months earlier and are less able than routinedata to provide data at intervals that permitcorrective action to be taken.Immunization117

Selection of IndicatorsThis chapter presents immunization coverage anddropout rates, which are used by many programsas indicators of the availability, accessibility, anduse of services, as well as of other programcharacteristics, and the vaccine wastage rate. Theindicators are listed below:• Proportion of infants born protected againstneonatal tetanus• BCG coverage• DTP1 coverage• DTP3 coverage• OPV3 coverage• Measles coverage• HepB3 coverage• Hib3 coverage• Dropout from DTP1 to DTP3• Fully immunized child• Vaccine wastage rateThese are the main indicators used bygovernments, organizations, donors, andcommunity leaders to monitor progress towardimmunization objectives and to make strategicdecisions on resource allocation. However, theindicators presented in this chapter do notconstitute a comprehensive set of indicators formonitoring and evaluating immunizationprograms. Coverage rates for DTP2, and the firstand second doses of OPV, HEPB, and HIB arenot included here but can easily be calculated byadapting the general definition of annualimmunization coverage (i.e., the proportion of thetarget population that has been vaccinated) andsurvey-based coverage (i.e., the proportion ofchildren 12-23 months who were immunized witha specific vaccine before 12 months of age).covered in this chapter. Examples of indicators ofquality include the number of immunizationsessions that are actually held compared to thenumber planned; vaccine usage; use of a sterilesyringe and needle for each injection; and parents’knowledge of common side effects and when toreturn for additional immunizations. Readers arereferred to Immunization Essentials: A PracticalField Guide for further details about these andother indicators of quality (USAID, 2003).Indicators for monitoring the information systemare also not addressed, but should measure theaccuracy as well as the completeness and timelinessof reporting. No one indicator can stand alone.A mix of indicators is necessary to obtain a morecomplete picture of services and to identifyproblems that should be investigated, as well aslikely solutions.The wording of some coverage indicators, such asDTP3 may be modified slightly to reflect the levelat which monitoring is implemented. For example,GAVI uses DTP3 coverage by district as anindicator of progress towards its goals whereasmany governments use national-level estimates tomeasure whether they are achieving theirobjectives. It should also be noted that thedescription of each coverage indicator includesinstructions on how to estimate the number ofsurviving children for use in calculating theindicator from routine data and a discussion ofthe relative strengths and limitations of routineandsurvey-based coverage estimates. While thisapproach may seem redundant compared tocombining all the coverage rates into one overallindicator, it ensures that the description of eachcoverage indicator is as complete as possible andconsistent with the format used in the rest of theguide.Many output measures and indicators of healthsystem performance that are used by managers toassess quality, efficiency, effort, and impact are not118Chapter 4

PROPOROPORTIONOF INFNFANTANTS BORNPROTECTEDAGAINSAINSTNEONEONATALAL TETETANUSDefinitionProportion of infants born protected againstneonatal tetanus in a specified period (usually 12months).Numerator: Number of infants born protectedagainst neonatal tetanus in a specified period.Denominator: Total number of live births inthe specified period.“Protection at birth:” For prevention of neonataland maternal tetanus, WHO recommends givingwomen a series of five doses of tetanus toxoid(TT). Each dose increases the level and protectionagainst tetanus. A woman who receives five dosesof tetanus toxoid is fully immunized throughouther childbearing years. If the mother has notreceived five lifetime TT doses, the child isconsidered “fully protected” against neonataltetanus if the mother received at least two TTdoses within the past three years, or three doseswithin the past five years, or four doses within thepast 10 years. Table 4.2 presents the tetanus toxoidimmunization schedule for women of childbearingage and pregnant women without previousexposure to tetanus toxoid-containing vaccines.Measurement ToolsRoutine administrative data; population-basedsurveys (e.g. MICS)What It MeasuresThis indicator measures the percent of birthsprotected against neonatal tetanus at the time ofdelivery among clients in a given program or inthe general population. While neonatal tetanuswas eliminated in industrialized countries as farback as the 1950s, it is still a major killer of infantsin the developing world, responsible for no lessthan 200,000 infant deaths every year andaccounting for 14% of all neonatal deaths. Up to70% of all babies that develop the disease die intheir first month of life. The goal of eliminatingmaternal and neonatal tetanus by 2005 wasdeclared by UNICEF, WHO, and UNFPA. Inthis context, elimination is defined as a rate ofneonatal tetanus below 1 per 1000 live births peryear at the district level (WHO, 2000d).How to Measure ItThe data requirements for calculating PABcoverage from routine administrative data are thenumber of doses of TT given to each pregnantwoman, the date each dose was given, and thenumber of children who received DTP1. Motherswho bring children for DTP1 immunization areasked for their immunization cards to determinehow many valid doses of tetanus toxoid have beenreceived. A dose is considered valid when theminimum required interval between doses hasbeen observed, as shown in Table 4.2. The periodof protection given by the number of doses themother has had is then determined and comparedwith the date of birth of the child. If the childwas born during the period of protection providedby the last valid dose, then the child is consideredprotected at birth. The number of childrenprotected at birth against neonatal tetanus iscumulated monthly and annually. PAB coverageis calculated from routine data by dividing thenumber of children protected at birth accordingto their mother’s TT vaccination history in aspecified period by the total number of childrenwho received DTP1 in that specified period(WHO, 1998).To estimate the extent of protection at birth inthe Multiple Indicator Cluster Surveys, all womenage 15-49 are first asked whether they have evergiven birth, and if so, the date of the last live birth,Immunization119

Table 4.2. Tetanus toxoid immunization schedule for women of childbearing age and pregnant womenwithout previous exposure to tetanus toxoid-containing vaccines (TT, Td, or DTP)Dose of TT,Td, or DTP12345When to giveAt first contact or as early as possible inpregnancyAt least four weeks after TT1At least six months after TT2 or duringsubsequent pregnancyAt least one year after TT3 or duringsubsequent pregnancyAt least one year after TT4 or duringsubsequent pregnancyExpected duration of protection*NoneOne to three yearsAt least five yearsAt least 10 yearsFor all childbearing years andpossibly longerNotes:(a) Increasing numbers of women have documentation of prior receipt of vaccines containing tetanus toxoid, e.g. inearly childhood or at school age. As the women reach childbearing age the incidence of maternal and neonataltetanus is expected to decline further: three properly spaced doses of DTP given in childhood are consideredequivalent in protection to two doses of TT/Td given in adulthood.(b) Recent studies suggest that the duration of protection may be longer than indicated in the table. This matter iscurrently under review.(c) Td – Tetanus-Diphtheria toxoid. Bivalent boosters given to children aged seven years and over and to adolescentsand adults with reduced diphtheria component to avoid reactions.Source: http://who.int/vaccines/en/tetanus.shtmland whether or not that child was alive at the timeof the interview. If the last child was born in theperiod 0-11 months before the survey, the motherwas then asked whether she had a vaccination cardor other documentary evidence of vaccination. Ifa card is presented to the interviewer, it is used toassist in obtaining information on the number oftetanus toxoid doses received while the womanwas pregnant with her last child. If the motherreports two or more doses during her lastpregnancy, the survey does not ascertain whethershe ever had earlier TT doses. If the mother120reports only one dose or none while she waspregnant with her last child, information is thencollected on earlier doses of TT (that is, dosesreceived during or before the next-to-lastpregnancy or between pregnancies). If the dateon which TT was administered is unknown, thewoman is asked to estimate how many years agoshe received the last dose of TT. In settings wherethe percentage of mothers having cards is low andthe precise dates of TT immunization areunknown, the indicator can be estimated usingboth card and history information.Chapter 4

Strengths and LimitationsionsThis indicator provides a means of monitoring TTcoverage through routine reporting that ispotentially more accurate than the “TT2+” methodin countries with high coverage of DTP1 vaccine.*The indicator can also be used to reduce missedopportunities for tetanus toxoid vaccination.Mothers of children classified as “not protected”can be vaccinated immediately, and at subsequentvisits if need be, to ensure that their next childwill be protected against neonatal tetanus.Variations in the definition of the numerator anddenominator for measuring protection at birthagainst neonatal tetanus may give rise todifferences in the magnitude and reliability of theindicator. For example, program statistics recordthe total number of doses of a vaccine in theprevious 12 months, whereas the MICS surveystend to record the number of doses given duringboth the current and earlier pregnancies and dosesadministered at times other than those specifiedin order to get at the number of lifetime doses.Some surveys such as the DHS report the numberof women who received at least two vaccinationsduring their last pregnancy in a reference periodthat may be up to five years. The MICS-approachof estimating protection at birth has the advantageof including doses of TT that were notadministered during the last pregnancy in thedefinition of the numerator.Routine administrative data have the disadvantagethat they may be incomplete or inaccurate (WHO,1999a). They are also subject to selection bias andare not representative of the general population,particularly when ANC coverage is low. However,they provide the only way of monitoring coverageon an annual basis and may be more reliable thanself-reported data.Surveys provide the only means of obtainingpopulation-based coverage, but because themother’s vaccination card may not be available tohelp the interviewer ascertain the number andtiming of doses received, surveys may tend to relyon self-reporting. Such reports are subject to recallbias that is likely to increase with the length ofthe recall period. Poor recall may result in adownward bias in estimates of the share of birthsprotected. In settings where receiving injectionsis common, respondents may also erroneouslyreport having received tetanus toxoid.* In the “TT2+” method, the numerator is calculated as thenumber of protective doses of TT (TT2, TT3, TT4 andTT5) given to pregnant women during a calendar year, andthe denominator as the estimated number of live birthsduring that calendar year. However, the sum of TT+ dosesunderestimates the number of protected pregnant women,since only pregnant women who received a TT dose duringtheir pregnancy are counted. Other pregnant womenexcluded from the numerator of the “TT2+” indicator mayhave already received TT5 and consequently were notvaccinated during pregnancy.Immunization121

BCG CBCG COVERAVERAGEDefinitionWhen calculating Bacille Calmette-Guerin(vaccine) (BCG) coverage from population-basedsurveys, the definition is as follows:Proportion of children age 12-23 months whowere immunized with BCG before 12 months ofage.Numerator: Number of children age 12-23months who were immunized with BCGbefore 12 months of age.Denominator: The number of children 12-23months of age surveyed.When calculating annual coverage of BCG fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withBCG in that calendar yearNumerator: Number immunized by 12 monthswith BCG in a specified calendar yearDenominator: Total number of survivinginfants less than 12 months of age in the sameyearMeasurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresThis indicator measures protection of childrenagainst miliary tuberculosis or tuberculousmeningitis. Tuberculosis is estimated to result in2.6 million deaths worldwide annually and 3.8million notified cases. Although the disease ismore common in adults, it is usually more seriousin infants, children, and adolescents. With thedeterioration of public health services in somecountries, and with the advent of HIV infection,the number of TB cases has escalated. Presently,WHO and UNICEF recommend that, forasymptomatic HIV-infected children living inareas where the risk of tuberculosis is high, BCGstill be given at birth or as soon as possiblethereafter in accordance with standard childhoodimmunization policies, but that it be withheld ininfants thought to have symptomatic HIVinfection.How to Measure ItWhen collecting data through household surveys,caretakers are asked to present the child’svaccination card to the interviewer so that it canbe used to record the actual dates at whichvaccinations were given. If a health card cannotbe presented, then all of the information aboutvaccination of children is collected from themother, based on her memory about whether ornot the child received those vaccinations.Interviewers may check the child for the BCGscar to validate the mother’s report. The BCGcoverage rate calculated on the basis of both cardand history information is recommended. Theindicator can be measured at both the nationaland sub-national/district levels.To determine the level of BCG coverage fromroutine data, two pieces of information are needed:The first is the number of doses of BCGadministered to children aged 0-11 months withina given calendar year (the numerator). The secondis the number of infants surviving to 12 monthsof age (live births – infant deaths) in the targetImmunization123

population in that specified calendar year. Thebox below illustrates how to calculate the annualcoverage rate for BCG vaccine.Calculating Annual Coverage for BCG Vaccine for 2003from Routine Administrative DataBCG coverageNumber immunized by 12 months with BCG in 2003 X 100Number of surviving infants

program purposes as specific areas within eachdistrict/subnational unit where special attentionand resources need to be diverted may beidentified. This is particularly relevant in thecontext of health sector reform anddecentralization. Routine data are moreappropriate, therefore, for active monitoring ofimmunization programs. However, caution needsto be exercised in the use of administrative dataon the number of doses administered due topotential problems with the completeness ofreporting (i.e., the proportion of sites submittingreports) and the degree to which data from theprivate sector are included in routine reports.Population-based surveys can provide moreaccurate and representative data for estimatingcoverage and can be helpful in validating theroutine reporting system. However, conducting asurvey is expensive and not done frequently.Furthermore, care should be taken when designingcluster surveys to ensure geographicalrepresentation and to avoid selection bias. Cardbaseddata increase the validity of survey estimates,but the vaccination card may not be widelyavailable.Immunization125

DTP1 TP1 CTP1 COVERAVERAGEDefinitionWhen calculating DTP1 coverage frompopulation-based surveys, the definition is asfollows:Proportion of children age 12-23 months whoreceived the first dose of Diphtheria, Tetanus, andPertussis (DTP1) vaccine before 12 months ofage.Numerator: Number of children age 12-23months who were immunized with first doseof DTP (DTP1) before 12 months of age.Denominator: The number of children 12-23months of age surveyed.When calculating annual coverage of DTP1 fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withDTP1 in that calendar year.Numerator: Number immunized by 12 monthswith DTP1 in a specified calendar year.Denominator: Total number of survivinginfants less than 12 months of age in the sameyear.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresThe indicator measures coverage of the first dosein a three-dose series of DTP and is commonlyused as an approximation of availability of, accessto, and initial use of immunization services bychildren. It reflects that the child has been incontact with a health worker and initiatedvaccination. The indicator assumes that there is afunctioning EPI program in place including healthproviders trained in EPI and adequate supplies tocarry out immunization services, and that thecommunity understands information regardingimmunization.How to Measure ItWhen collecting data through household surveys,caretakers are asked to present the child’svaccination card to the interviewer so that it canbe used to record the actual dates at whichvaccinations were given. If a health card cannotbe presented, then all of the information aboutthe vaccination of children is collected from themother, based on her memory about whether ornot the child received those vaccinations. For DTP(and polio) vaccines, follow-up questions are askedabout the number of times the child received thevaccine. The DTP1 coverage rate calculated onthe basis of both card and history information isrecommended. The indicator can be measured atthe national and sub-national/district levels.DTP1 coverage by age 12 months can also becalculated from routine data. Daily tallies ofvaccinations are added at the end of the monthand converted into a percentage as a measure ofcoverage. The data requirements for calculatingDTP1 coverage from routine data are the numberof infants immunized by age 12 months withDTP1 vaccine in the specified calendar year(numerator) and the number of infants survivingto 12 months of age (live births – infant deaths)in the target population in that specified calendarImmunization127

year. The box below illustrates how to calculatethe annual coverage rate for DTP1 vaccine.Calculating Annual Coverage for DTP1 Vaccine for2003 from Routine Administrative DataDTP1 coverageNumber immunized by 12 months with DTP1 in 2003 X 100Number of surviving infants

Related indicators are:• Proportion of districts/subnational unitsreporting < 50% DTP1 coverage• Proportion of districts/subnational unitsreporting 50-79% DTP1 coverage• Proportion of districts/subnational unitsreporting ≥ 80% DTP1 coverageThese indicators measure the performance ofdifferent units at the subnational level againstexpected national targets and are easily calculatedfrom administrative data monitoring the numberof doses administered to the target population.Population-based surveys can provide moreaccurate and representative data for estimatingDTP coverage and can be helpful in validatingthe routine reporting system. However,conducting a survey is expensive and not donefrequently. Furthermore, care should be takenwhen designing cluster surveys to ensuregeographical representation and to avoid selectionbias. Card-based data increase the validity ofsurvey estimates, but the vaccination card may notbe widely available.Estimates based on routine data are relativelyinexpensive to collect. Unlike surveys, routine datacan also be analyzed monthly and are available atany administrative level. The use of routineadministrative data for calculating indicators overtime will provide information on trends and thusguide management decisions. Furthermore,disaggregation of the routine-based estimates bydistricts/subnational units serves programpurposes as specific areas within each district/subnational unit where special attention andresources need to be diverted may be identified.This is particularly relevant in the context of thehealth sector reform and decentralization currentlyunderway. Routine data are more appropriate,therefore, for active monitoring of immunizationprograms. However, caution needs to be exercisedin the use of administrative data on the numberof doses administered due to potential problemswith the completeness of reporting (i.e., theproportion of sites submitting reports) and thedegree to which data from the private sector areincluded in routine reports.Immunization129

DTP3 TP3 CTP3 COVERAVERAGEDefinitionWhen calculating DTP3 coverage frompopulation-based surveys, the definition is asfollows:Proportion of children age 12-23 months whoreceived three doses of DTP (DTP3) vaccine byage 12 months.Numerator: Number of children age 12-23months who received three doses of DTPvaccine by age 12 months.Denominator: Total number of children age12-23 months surveyed.When calculating annual coverage of DTP3 fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withDTP3 in that calendar year.Numerator: Number immunized by 12 monthswith DTP3 in a specified calendar year.Denominator: Total number of survivinginfants less than 12 months of age in the sameyear.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresThis indicator measures the ability of the healthsystem to deliver a series of vaccinations. Itindicates continuity of use of immunizationservices by caretakers and client satisfaction withservices. Hence, it measures the effectiveness ofroutine service delivery.ImmunizationHow to Measure ItWhen collecting data through household surveys,caretakers are asked to present the child’svaccination card to the interviewer so that it canbe used to record the actual dates at whichvaccinations were given. If a health card cannotbe presented, then all of the information aboutvaccination of children is collected from themother, based on her memory about whether ornot the child received those vaccinations. For DTP(and polio) vaccines, follow-up questions are askedabout the number of times the child received thevaccine. The DTP3 coverage rate calculated onthe basis of both card and history information isrecommended. The indicator can be measured atthe national and sub-national/district levels.DTP3 coverage by age 12 months can also becalculated from routine data. Daily tallies ofvaccinations are added at the end of the monthand converted into a percentage as a measure ofcoverage. The data requirements for calculatingDTP3 coverage from routine data are the numberof infants immunized by age 12 months withDTP3 vaccine in the specified calendar year(numerator) and the number of infants survivingto 12 months of age (live births – infant deaths) inthe target population in that specified calendaryear. The box below illustrates how to calculatethe annual coverage rate for DTP3 vaccine.Calculating Annual Coverage for DTP3 Vaccine for2003 from Routine Administrative DataDTP3 coverageNumber immunized by 12 months with DTP3 in 2003 X 100Number of surviving infants

The difficulty of calculating and interpretingimmunization coverage rates from routine dataoften stems from problems related to estimatingthe size of the target population. These problems,in turn, are the result of inaccurate or outdatedcensus counts, population migrations, andunforeseen changes in birth rates or infantmortality. Where data on the numbers of survivinginfants are unavailable, the total number of livebirths can be estimated from the total populationand crude birth rate and infant mortality rate in ageographic area as follows:132Number of surviving infants = Total populationx crude birth rate x (1 - IMR)For example:Estimating the Number of Surviving Infants: ExampleTotal population: 5,500,000Crude birth rate (CBR): 30/1000Infant mortality rate (IMR): 80/1000Number of surviving infants =Total population x CBR x (1 - IMR)= 5,500,000 x 30/1000 x (1 - 0.080)= 5,500,000 x 0.030 x 0.920= 151,800Source: Immunization Essentials: A Practical FieldGuide (USAID, 2003).If the above population figures cannot be obtainedfrom a national or community census and if nodata are available on the crude birth rate or infantmortality rate, the number of children under oneyear in the population can be estimated bymultiplying the total population by 4% (WHO,2002b). For example, if the total population is30,000, then the number of children under oneyear is 30,000 x 4/100 = 1200. If a more precisepercentage of children under one year in thepopulation is known, this estimate should be usedto calculate the annual target population.As with other EPI vaccines monitoring of DTP3immunization coverage, based on doses of vaccineadministered in the target age group (usuallyinfants < 1 year of age) should be done on amonthly basis at the health facility and districtlevel (WHO, 1996). This requires estimating themonthly target population. The monthly targetpopulation can be estimated by dividing thenumber of children under one year of age by 12.(If the annual target population of children underone year of age is 1200, the monthly target is 1200/12 = 100).One of the milestones established by GAVI is thatby 2010 or sooner all countries will have routineimmunization coverage at 90% nationally with atleast 80% coverage in every district. The cut-offDTP3 coverage can be adjusted depending on thelocal situation/progress and what a countryconsiders “high” or “low” coverage (WHO,2002b).Strengths and LimitationsIn countries with established immunizationprograms, the DTP3 coverage rate can be used asa proxy for full immunization coverage. However,the indicator is a crude measure in that it does nottake into account the timeliness of each of thedoses of DTP or the interval between doses.Related indicators are:• Proportion of districts/subnational unitsreporting < 50% DTP3 coverage• Proportion of districts/subnational unitsreporting 50-79% DTP3 coverage• Proportion of districts/subnational unitsreporting ≥ 80% DTP3 coverageThese indicators conform to GAVI requirements,and measure the performance of different units atthe subnational level against expected nationaltargets. The indicators are easily calculated fromadministrative data monitoring the number ofdoses administered to the target population.Chapter 4

Estimates based on routine data are relativelyinexpensive to collect. Unlike surveys, routine datacan also be analyzed monthly and are available atany administrative level. The use of routineadministrative data for calculating indicators overtime will provide information on trends and thusguide management decisions. Furthermore,disaggregation of the routine-based estimates bydistricts/subnational units serves programpurposes as specific areas within each district/subnational unit where special attention andresources need to be diverted may be identified.This is particularly relevant in the context of healthsector reform and decentralization. Routine dataare more appropriate, therefore, for activemonitoring of immunization programs. However,caution needs to be exercised in the use ofadministrative data on the number of dosesadministered due to potential problems with thecompleteness of reporting (i.e., the proportion ofsites submitting reports) and the degree to whichdata from the private sector are included in routinereports.Population-based surveys can provide moreaccurate and representative data for estimatingDTP coverage and can be helpful in validatingthe routine reporting system. However,conducting a survey is expensive and not donefrequently. Furthermore, care should be takenwhen designing cluster surveys to ensuregeographical representation and to avoid selectionbias. Card-based data increase the validity ofsurvey estimates, but the vaccination card may notbe widely available.Immunization133

OPV3 COPV3 COVERAVERAGEDefinitionProportion of children age 12-23 months whoreceived three doses of oral polio vaccine (OPV3)by age 12 months.Numerator: Number of children age 12-23months who received three doses of OPV byage 12 months.Denominator: Total number of children age12-23 months surveyed.When calculating annual coverage of OPV3 fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withOPV3 in that calendar year.Numerator: Number immunized by 12 monthswith OPV3 in a specified calendar year.Denominator: Total number of survivinginfants less than 12 months of age in the sameyear.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresThis indicator measures the ability of the healthsystem to deliver a series of vaccinations. Itindicates continuity of use of immunizationservices by caretakers, and client satisfaction withservices. Hence, it measures the effectiveness ofroutine service.How to Measure ItWhen collecting data through household surveys,caretakers are asked to present the child’svaccination card to the interviewer so that it canbe used to record the actual dates at whichvaccinations were given. If a dose of OPV is givenat a health facility within the first two weeks oflife (this dose being referred to as OPV0), this isnot counted as one of the three doses required tocalculate the numerator for this indicator. If ahealth card cannot be presented, then all of theinformation about the vaccination of children iscollected from the mother, based on her memoryabout whether or not the child received thosevaccinations. Follow-up questions are asked aboutthe number of times the child received the vaccine.To calculate this indicator, campaign-administereddoses should be excluded. These doses do notreflect the utilization of routine immunizationservices. For that reason, survey-based questionsmust indicate whether OPV doses were receivedthrough the routine system or campaigns. TheOPV3 coverage rate calculated on the basis of bothcard and history information is recommended.The indicator can be measured at the national andsub-national/district levels.OPV3 coverage by age 12 months can also becalculated from routine data. Daily tallies ofvaccinations are added at the end of the monthand converted into a percentage as a measure ofcoverage. The data requirements for calculatingOPV3 coverage from routine data are the numberof infants immunized by age 12 months withOPV3 vaccine in the specified calendar year(numerator) and the number of infants survivingto 12 months of age (live births – infant deaths)Immunization135

in the target population in that specified calendaryear. The box below illustrates how to calculatethe annual coverage rate for OPV3 vaccine.Calculating Annual Coverage for OPV3 Vaccine for2003 from Routine Administrative DataOPV3 coverageNumber immunized by 12 months with OPV3 in 2003 X 100Number of surviving infants < 12 months of age in 2003The difficulty of calculating and interpretingimmunization coverage rates from routine dataoften stems from problems related to estimatingthe size of the target population. These problems,in turn, are the result of inaccurate or outdatedcensus counts, population migrations, andunforeseen changes in birth rates or infantmortality. Where data on the numbers of survivinginfants are unavailable, the total number of livebirths can be estimated from the total populationand crude birth rate and infant mortality rate in ageographic area as follows:Number of surviving infants = Total populationx crude birth rate x (1 - IMR)For example:If the population figures presented in the previouscolumn cannot be obtained from a national orEstimating the Number of Surviving Infants: ExampleTotal population: 5,500,000Crude birth rate (CBR): 30/1000Infant mortality rate (IMR): 80/1000Number of surviving infants =Total population x CBR x (1 - IMR)= 5,500,000 x 30/1000 x (1 - 0.080)= 5,500,000 x 0.030 x 0.920= 151,800Source: Immunization Essentials: A Practical FieldGuide (USAID, 2003).community census and if no data are available onthe crude birth rate or infant mortality rate, thenumber of children under one year in thepopulation can be estimated by multiplying thetotal population by 4% (WHO, 2002b). Forexample, if the total population is 30,000, thenthe number of children under one year is 30,000 x4/100 = 1200. If a more precise percentage ofchildren under one year in the population isknown, this estimate should be used to calculatethe annual target population.As with other EPI vaccines, monitoring of OPV3immunization coverage, based on doses of thevaccine administered in the target age group(usually infants < 1 year of age), should be doneon a monthly basis at the health facility and districtlevel (WHO, 1996). This requires estimating themonthly target population. The monthly targetpopulation can be estimated by dividing thenumber of children under one year of age by 12.(If the annual target population of children underone year of age is 1200, the monthly target is 1200/12 = 100).One of the milestones established by GAVI is thatby 2010 or sooner all countries will have routineimmunization coverage at 90% nationally with atleast 80% coverage in every district. The cut-offOPV3 coverage can be adjusted depending on thelocal situation/progress and what a countryconsiders “high” or “low” coverage (WHO,2002b).Strengths and LimitationsEstimates based on routine data are relativelyinexpensive to collect. Unlike surveys, routine datacan also be analyzed monthly and are available atany administrative level. The use of routineadministrative data for calculating indicators overtime will provide information on trends and thusguide management decisions. Furthermore,disaggregation of the routine-based estimates bydistricts/subnational units serves programpurposes as specific areas within each district/subnational unit where special attention and136Chapter 4

esources need to be diverted may be identified.This is particularly relevant with the health sectorreform and decentralization currently underway.Routine data are more appropriate, therefore, foractive monitoring of immunization programs.However, caution needs to be exercised in the useof routine data on the number of dosesadministered due to potential problems with thecompleteness of reporting (i.e., the proportion ofsites submitting reports) and the degree to whichdata from the private sector are included in routinereports.Population-based surveys can provide moreaccurate and representative data for estimatingOPV3 coverage and can be helpful in validatingthe routine reporting system. However,conducting a survey is expensive and not donefrequently. Furthermore, care should be takenwhen designing cluster surveys to ensuregeographical representation and to avoid selectionbias. Card-based data increase the validity ofsurvey estimates, but the vaccination card may notbe widely available.Immunization137

MEASLESCOVERAVERAGEDefinitionWhen calculating measles coverage frompopulation-based surveys, the definition is asfollows:Proportion of children age 12-23 months whowere immunized with measles vaccine before age12 months.Numerator: Number of children age 12-23months who were immunized with measlesvaccine before age 12 months.Denominator: Total number of children age12-23 months surveyed.When calculating annual measles coverage fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withmeasles vaccine in that calendar year.Numerator: Number immunized by 12 monthswith measles vaccine in a given year.Denominator: Total number of survivinginfants less than 12 months of age in the sameyear.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresThis indicator measures protection againstmeasles, a disease of major public healthimportance. Measles remains one of the leadingcauses of child mortality in developing countriesand causes approximately 10% of all deaths amongchildren aged less than five years (WHO, 1994).In combination with disease surveillance data, thisindicator measures progress towards measlescontrol and elimination. The priorities forcountries pursuing accelerated measles controlinclude improving routine vaccination coveragelevels to at least 80% in all districts of everycountry, and achieving at least 90% coveragenationwide. Priorities for countries and regionswith a measles elimination goal include improvingroutine vaccination coverage levels to at least 90%in all districts of every country, resulting innationwide coverage greater than or equal to 95%(CDC, 1999).How to Measure ItWhen collecting data through household surveys,caretakers are asked to present the child’svaccination card to the interviewer so that it canbe used to record the actual dates at whichvaccinations were given. If a health card cannotbe presented, then all of the information aboutvaccination of children is collected from themother, based on her memory about whether ornot the child received those vaccinations. In somesituations, children may have received measlesdoses during campaigns. To calculate thisindicator, campaign-administered doses should beexcluded. These doses do not reflect the utilizationof routine immunization services. For that reason,survey-based questions must inquire as to whetherthe measles dose was received through the routinesystem or campaigns. The measles coverage ratecalculated on the basis of both card and historyinformation is recommended. The indicator canbe measured at both the national and sub-national/district levels.Measles-containing vaccine (MCV) coverage byage 12 months can also be calculated from routinedata. Daily tallies of vaccinations are added at theImmunization139

end of the month and converted into a percentageas a measure of coverage. The data requirementsfor calculating measles vaccine coverage fromroutine data are the number of infants immunizedby 12 months with measles vaccine in the specifiedcalendar year (numerator) and the number ofinfants surviving to 12 months of age (live births– infant deaths) in the target population in thatspecified calendar year. Measurement of thenumerator should be taken with caution so as notto include children over one year of age. The boxbelow illustrates how to calculate the annualcoverage rate for measles vaccine.Calculating Annual Coverage for Measles Vaccine for2003 from Routine Administrative DataMeasles coverageNumber immunized by 12 months with measles vaccine in 2003 X 100Number of surviving infants < 12 months of age in 2003The difficulty of calculating and interpretingimmunization coverage rates from routine dataoften stems from problems related to estimatingthe size of the target population. These, in turnare the result of inaccurate or outdated censuscounts, population migrations, and unforeseenchanges in birth rates or infant mortality. Wheredata on the numbers of surviving infants areunavailable, the total number of live births can beestimated from the total population and crudebirth rate and infant mortality rate in a geographicarea as follows:Number of surviving infants = Total populationx crude birth rate x (1 - IMR)For example:Estimating the Number of Surviving Infants: ExampleTotal population: 5,500,000Crude birth rate (CBR): 30/1000Infant mortality rate (IMR): 80/1000Number of surviving infants =Total population x CBR x (1 - IMR)= 5,500,000 x 30/1000 x (1 - 0.080)= 5,500,000 x 0.030 x 0.920= 151,800Source: Immunization Essentials: A Practical FieldGuide (USAID, 2003).If the above population figures cannot be obtainedfrom a national or community census and if nodata are available on the crude birth rate or infantmortality rate, the number of children under oneyear in the population can be estimated bymultiplying the total population by 4% (WHO,2002b). For example, if the total population is30,000, then the number of children under oneyear is 30,000 x 4/100 = 1200. If a more precisepercentage of children under one year in thepopulation is known, this estimate should be usedto calculate the annual target population.As with other EPI vaccines, monitoring of measlesimmunization coverage based on doses of thevaccine administered in the target age group(usually infants < 1 year of age) should be done ona monthly basis at the health facility and districtlevel (WHO, 1996). This requires estimating themonthly target population. The monthly targetpopulation can be estimated by dividing thenumber of children under one year of age by 12.(If the annual target population of children lessthan one year of age is 1200, the monthly target is1200/12 = 100).140Chapter 4

One of the milestones established by GAVI is thatby 2010 or sooner all countries will have routineimmunization coverage at 90% nationally with atleast 80% coverage in every district. It should benoted that even with very high immunizationcoverage (95%), susceptibles will continue toaccumulate fairly rapidly as the measles vaccine isnot 100% effective (WHO, 1996).Strengths and LimitationsEstimates based on routine data are relativelyinexpensive to collect. Unlike surveys, routine datacan also be analyzed monthly and are available atany administrative level. The use of routineadministrative data for calculating indicators overtime will provide information on trends and thusguide management decisions. Furthermore,disaggregation of the routine-based estimates bydistricts/subnational units serves programpurposes as specific areas within each district/subnational unit where special attention andresources need to be diverted may be identified.This is particularly relevant in the context of healthsector reform and decentralization. Routine dataare more appropriate, therefore, for activemonitoring of immunization programs.inflating coverage among the infant population.This may be because the tally and reporting formsdo not include a column for recordingimmunizations received by children older than 12months of age. Furthermore, supplementary dosesmay be confused with routine doses whenmeasuring this indicator.Population-based surveys provide more accurateand representative data for estimating coverageand can be helpful in validating the routinereporting system. However, conducting a surveyis expensive and not done frequently.Furthermore, care should be taken when designingcluster surveys to ensure geographicalrepresentation and to avoid selection bias. Cardbaseddata increase the validity of survey estimates,but the vaccination card may not be widelyavailable.Caution needs to be exercised in the use ofadministrative data on the number of dosesadministered due to potential problems with thecompleteness of reporting (i.e., the proportion ofsites submitting reports) and the degree to whichdata from the private sector are included in routinereports.This indicator does not indicate the capability ofthe health system to deliver a series of vaccines. Italso does not take into account the timeliness ofmeasles vaccination. Immunization for measlesshould be given as soon as possible after the childcompletes nine months of life to avoid the risk ofinfection by the wild measles virus. The numeratorincludes children who may have received thevaccine before the age of 9 months, thus conferringuncertain immunity. The numerator may alsoinclude children older than 12 months of age, thusImmunization141

HEPB3 CHEPB3 COVERAVERAGEDefinitionWhen calculating HepB3 coverage frompopulation-based surveys, the definition is asfollows:Proportion of children age 12-23 months whoreceived three doses of Hepatitis B (HepB3)vaccine by age 12 months.Numerator: Number of children age 12-23months who received three doses of HepBvaccine by age 12 months.Denominator: Total number of children age12-23 months surveyed.When calculating annual HepB3 coverage fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withHepB3 in that calendar year.Numerator: Number of children immunizedby 12 months with HepB3 in a given year.Denominator: Total number of survivinginfants less than 12 months of age in the sameyear.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresUniversal infant immunization against HepatitisB is now recognized as the proper strategy forevery country for the long-term control of chronicHBV infection and its sequelae (cirrhosis and livercancer). If the vaccine is administered beforeinfection, it prevents the development of theImmunizationdisease and the carrier state in almost allindividuals. On a population basis, HepB is mosteffective when used routinely as part of the infantimmunization schedule, although it can be usedin persons of any age. HepB vaccine schedulesare very flexible and the vaccine can be added tocoincide with DTP schedules so that additionalvisits for HepB vaccination would not be necessary.In countries with high perinatal transmission,babies are frequently infected at the time of birth.In those countries with the ability to reachnewborns, Hepatitis B vaccine should be offeredas soon as possible after birth, preferably withinthe first 24 hours of life. As WHO recommends,the routine vaccination of all infants as an integralpart of national immunization schedules shouldbe the highest priority in all countries. In countriesof high disease endemicity (HepB surface antigen(HbsAg) prevalence [8% or more]), routine infantHepB vaccination can rapidly reduce transmissionbecause most chronic infections are acquired as aresult of spread either from mother to baby or fromchild to child in the first year of life (WHO, 2003).The indicator measures HepB coverage. Theindicator is applicable to countries where routineinfant HepB vaccination has been introduced. Itassumes that there is a functioning EPI programin place including health providers trained inimmunization and adequate supplies to carry outEPI services, and that the community understandsinformation regarding immunization.How to Measure ItWhen collecting data through household surveys,and in countries where HepB vaccination isprovided with DTP vaccination, HepB3 coveragemay be calculated based on a child’s vaccinationcard and an interview with his/her caretaker.During the survey, caretakers are asked to present143

the child’s vaccination card to the interviewer sothat it can be used to record the actual dates atwhich vaccinations were given. If a health cardcannot be presented, then all of the informationabout vaccination of children is collected from themother, based on her memory about whether ornot the child received those vaccinations. Similarto DTP (and polio) vaccines, follow-up questionscan be asked about the number of times the childreceived the HepB vaccine.The HepB3 coverage rate calculated on the basisof both card and history information isrecommended. The indicator can be measured atboth the national and sub-national/district levels.In countries where a high proportion of chronicinfections are acquired perinatally (e.g. South EastAsia), a birth dose of Hepatitis B vaccine may begiven to infants when feasible. If a dose of HepBis given at a health facility within first two weeksof life, this dose is referred to as HepB1, and iscounted as one of the three doses required tocalculate the numerator for this indicator.HepB3 coverage by age 12 months can also becalculated from routine data. Daily tallies ofvaccinations are added at the end of the monthand converted into a percentage as a measure ofcoverage. The data requirements for calculatingHepB3 coverage from routine data are the numberof infants immunized by 12 months with HepB3vaccine in the specified calendar year (numerator)and the number of infants surviving to 12 monthsof age (live births – infant deaths) in the targetpopulation in that specified calendar year. Thebox below illustrates how to calculate the annualcoverage rate for HepB3 vaccine.Calculating Annual Coverage for HepB3 Vaccine for2003 from Routine Administrative DataHepB3 coverageNumber immunized by 12 months with HepB3 in 2003 X 100Number of surviving infants < 12 months of age in 2003144The difficulty of calculating and interpretingimmunization coverage rates from routine dataoften stems from problems related to estimatingthe size of the target population. These, in turn,are the result of inaccurate or outdated censuscounts, population migrations, and unforeseenchanges in birth rates or infant mortality. Wheredata on the numbers of surviving infants areunavailable, the total number of live births can beestimated from the total population and crudebirth rate and infant mortality rate in a geographicarea as follows:Number of surviving infants = Total populationx crude birth rate x (1 - IMR)For example:Estimating the Number of Surviving Infants: ExampleTotal population: 5,500,000Crude birth rate (CBR): 30/1000Infant mortality rate (IMR): 80/1000Number of surviving infants =Total population x CBR x (1 - IMR)= 5,500,000 x 30/1000 x (1 - 0.080)= 5,500,000 x 0.030 x 0.920= 151,800Source: Immunization Essentials: A Practical FieldGuide (USAID, 2003).If the above population figures cannot be obtainedfrom official census data or a community censusand if no data are available on the crude birth rateor infant mortality rate, the number of childrenunder one year in the population can be estimatedby multiplying the total population by 4% (WHO,2002b). For example, if the total population is30,000, then the number of children under oneyear is 30,000 x 4/100 = 1200. If a more precisepercentage of children under one year in thepopulation is known, this estimate should be usedto calculate the annual target population.Chapter 4

One of the milestones established by GAVI is thatby 2010 or sooner all countries will have routineimmunization coverage at 90% nationally with atleast 80% coverage in every district. As manycountries are at the early stages of the integrationof Hepatitis B vaccine into their routineimmunization programs, this indicator willtypically remain low for some time.Strengths and LimitationsThis indicator can be easily calculated fromadministrative data monitoring the number ofdoses administered to the target population.Estimates based on routine data are relativelyinexpensive to collect. Unlike surveys, routine datacan also be analyzed monthly and are available atany administrative level. The use of routineadministrative data for calculating indicators overtime will provide information on trends and thusguide management decisions. Furthermore,disaggregation of the routine-based estimates bydistricts/subnational units serves programpurposes as specific areas within each district/subnational unit where special attention andresources need to be diverted may be identified.This is particularly relevant in the context of thehealth sector reform and decentralization currentlyunderway. Routine data are more appropriate,therefore, for active monitoring of immunizationprograms. However, caution needs to be exercisedin the use of administrative data on the numberof doses administered due to potential problemswith the completeness of reporting (i.e., theproportion of sites submitting reports) and thedegree to which data from the private sector areincluded in routine reports.Population-based surveys provide more accurateand representative data for estimating coverageand can be helpful in validating the routinereporting system. However, conducting a surveyis expensive and not done frequently.Furthermore, care should be taken when designingcluster surveys to ensure geographicalrepresentation and to avoid selection bias. Cardbaseddata increase the validity of survey estimates,but the vaccination card may not be widelyavailable.Immunization145

HIB3 CHIB3 COVERAVERAGEDefinitionWhen calculating Hib3 coverage frompopulation-based surveys, the definition is asfollows:Proportion of children age 12-23 months whoreceived three doses of Haemophilus InfluenzaeType B (Hib3) vaccine by age 12 months.Numerator: Number of children age 12-23months who received three doses of Hibvaccine by age 12 months.Denominator: Total number of children age12-23 months surveyed.When calculating annual Hib3 coverage fromroutine data, the definition is as follows:Proportion of infants 0-11 months of age in aspecified calendar year who were immunized withHib3 in that calendar year.Numerator: Number of children immunizedby 12 months with Hib3 in a given year.Denominator: Total number of survivinginfants less than 12 months of age in the sameyear.Measurement ToolPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresSince 1998, WHO has recommended that Hibvaccine be included in routine infantimmunization programs in all countries whereresources permit its use and the burden of diseaseis established. One of the milestones set by theGlobal Alliance for Vaccines and Immunizationis that by 2005, 50% of the poorest countries withhigh disease burden and adequate delivery systemswill have introduced Hib vaccine (GAVI, 2001).Where it has been studied carefully, Hib istypically the leading cause of bacterial meningitisin infants and children less than five years old andaccounts for one-third to one-half of all cases ofbacterial meningitis in this age group. Studies havealso shown that Hib accounts for up to one-quarterof the severe pneumonia cases in young children.WHO (2000c) estimates that without vaccination,400,000 children die each year of Hib disease.Infants 0-11 months of age require a primary doseschedule of three doses of Hib conjugate vaccinein the first year of life. This indicator measuresHib3 coverage and the impact of the integrationof Hib vaccine into the routine nationalimmunization program.How to Measure ItIn countries that are newly integrating Hibvaccination into national immunization programs,Hib3 coverage by age 12 months should becalculated from routine data. Daily tallies ofvaccinations are added at the end of the monthand converted into a percentage as a measure ofcoverage.Hib3 coverage by age 12 months can also becalculated from routine data. Daily tallies ofvaccinations are added at the end of the monthand converted into a percentage as a measure ofcoverage. The data requirements for calculatingHib3 coverage from routine data are the numberof infants immunized by 12 months with Hib3vaccine in the specified calendar year (numerator)and the number of infants surviving to 12 monthsof age (live births – infant deaths) in the targetpopulation in that specified calendar year. TheImmunization147

ox below illustrates how to calculate the annualcoverage rate for Hib3 vaccine.Calculating Annual Coverage for Hib3 Vaccine for 2003from Routine Administrative DataHib3 coverageNumber immunized by 12 months with Hib3 in 2003 X 100Number of surviving infants < 12 months of age in 2003When using administrative data, caution must alsobe taken not to include children aged 12 monthsand older in the calculation of coverage rates.The difficulty of calculating and interpretingimmunization coverage rates from routine data oftenstems from problems related to estimating the sizeof the target population. These, in turn, are the resultof inaccurate or outdated census counts, populationmigrations, and unforeseen changes in birth rates orinfant mortality. Where data on the numbers ofsurviving infants are unavailable, the total numberof live births can be estimated from the totalpopulation and crude birth rate and infant mortalityrate in a geographic area as follows:148Number of surviving infants = Total populationx crude birth rate x (1 - IMR)For example:Estimating the Number of Surviving Infants: ExampleTotal population: 5,500,000Crude birth rate (CBR): 30/1000Infant mortality rate (IMR): 80/1000Number of surviving infants =Total population x CBR x (1 - IMR)= 5,500,000 x 30/1000 x (1 - 0.080)= 5,500,000 x 0.030 x 0.920= 151,800Source: Immunization Essentials: A Practical FieldGuide (USAID, 2003).If the population figures presented in the previouscolumn cannot be obtained from official censusdata or a community census and if no data areavailable on the crude birth rate or infant mortalityrate, the number of children under one year in thepopulation can be estimated by multiplying thetotal population by 4% (WHO, 2002b). Forexample, if the total population is 30,000, thenthe number of children under one year is 30,000 x4/100 = 1200. If a more precise percentage ofchildren under one year in the population isknown, this estimate should be used to calculatethe annual target population.In countries in which Hib vaccine has beenadministered to infants for a number of years, theindicator can also be calculated using householdsurvey data. When using survey data, the indicatoris defined as the proportion of children age 12-23months who received three doses of Hib vaccineby age 12 months. The numerator is defined asthe number of children age 12-23 months whoreceived three doses of Hib vaccine by age 12months. The denominator for survey-basedestimates is the total number of children age 12-23 months surveyed.When collecting data through household surveys,caretakers are asked to present the child’svaccination card to the interviewer so that it canbe used to record the actual dates at whichvaccinations were given. If a health card cannotbe presented, then all of the information aboutvaccination of children is collected from themother, based on her memory about whether ornot the child received Hib3 vaccinations. Similarto DTP (and polio) vaccines, follow-up questionscan be asked about the number of times the childreceived the Hib3 vaccine. This should includeonly vaccinations that the child has receivedthrough the routine system, and not those receivedduring any national Hib campaigns that may havebeen conducted. The Hib3 coverage ratecalculated on the basis of both card and historyinformation is recommended. The indicator canbe measured at the national and sub-national/district levels.Chapter 4

One of the milestones established by GAVI is thatby 2010 or sooner all countries will have routineimmunization coverage at 90% nationally with atleast 80% coverage in every district. As manycountries are at the early stages of the integrationof Hib vaccine into their routine immunizationprograms, this indicator will typically remain lowfor some time.Strengths and LimitationsThis indicator can be easily calculated fromadministrative data monitoring the number ofdoses administered to the target population.Estimates based on routine data are relativelyinexpensive to collect. Unlike surveys, routine datacan also be analyzed monthly and are available atany administrative level. The use of routineadministrative data for calculating indicators overtime will provide information on trends and thusguide management decisions. Furthermore,disaggregation of the routine-based estimates bydistricts/subnational units serves programpurposes as specific areas within each district/subnational unit where special attention andresources need to be diverted may be identified.This is particularly relevant with the health sectorreform and decentralization currently underway.Routine data are more appropriate, therefore, foractive monitoring of immunization programs.cluster surveys to ensure geographicalrepresentation and to avoid selection bias.Estimates based on mother’s recall may becomplicated by the fact that Hib conjugatevaccines are available as a monovalent vaccine (Hibconjugate vaccine only) or in combination withother routine vaccines (e.g., DTP, DTP-HepatitisB). It may be difficult for mothers to knowwhether a child received Hib vaccine if acombination vaccine is used, especially in the earlyphases of the introduction of Hib vaccine into theroutine national immunization program. Anotherlimitation with estimates based on mother’s recallis low recognition of this causative agent combinedwith the fact that most local languages do not havea specific name for the disease. Hib can causemeningitis and pneumonia, but it is only one causeof those diseases. This is unlike the case withmeasles, pertussis, and polio, which are allrecognizable distinct diseases in the vernacular.Card-based data increase the validity of surveyestimates, but the vaccination card may not bewidely available. Note that this indicator does nottake into account the timeliness of Hibvaccination.Caution needs to be exercised in the use ofadministrative data on the number of dosesadministered due to potential problems with thecompleteness of reporting (i.e., the proportion ofsites submitting reports) and the degree to whichdata from the private sector are included in routinereports. As a result, coverage estimated throughthis indicator may be less accurate in places wheredata quality is poor.Population-based surveys provide more accurateand representative data for estimating coverageand can be helpful in validating the routinereporting system. However, conducting a surveyis expensive and not done frequently.Furthermore, care should be taken when designingImmunization149

DROPOUTFROMOMOM DTP1TO DTP3DefinitionSurvey dataProportion of children age 12-23 months whoreceived DTP1 but did not receive DTP3immunization before age 12 months.Numerator: Number of children 12-23 monthswho received DTP1 minus number of children12-23 months who had not received DTP3immunization by their first birthday.Denominator: Number of children 12-23months who received DTP1 before age 12months.Routine dataDTP1-DTP3 dropout rateNumerator: Cumulative total of DTP1 minuscumulative total of DTP3 immunizationsgiven to children below the age of one yearduring a specified reference period (forexample, during the past 12 months).Denominator: Cumulative total of DTP1 givento children below the age of one year duringthe same period.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS); routine administrative dataWhat It MeasuresThis indicator measures the continuation ofimmunization and perceived quality of services,that is, whether children were in contact with EPIservices but dropped out of the system before theseries was completed. Dropout rates based onroutine administrative data also provide a usefulsummary measure of the overall performance ofroutine immunization services and may reflect thequality of communication between parents andhealth workers.How to Measure ItWhen using surveys, the data requirements formeasuring this indicator are the number ofchildren age 12-23 months who received DTP1vaccine and the number who received DTP3vaccine before the age of 12 months. Based onthis information, one can calculate the proportionof children who dropped out of the system. It isrecommended that dropout rates be based on bothcard and history information. DTP1-DTP3dropout rates are usually calculated from routineadministrative data as the difference in thecumulative totals of DTP1 and DTP3 dosesadministered to children below the age of one yearin a given reference period (for example, duringthe last 12 months), divided by the cumulativenumber of doses of DTP1 administered duringthe same period. Dropout rates can be calculatedusing either absolute numbers or percentages.The quality of utilization is considered good ifthe dropout rate in the target age group is lessthan 10 percent and poor if the dropout rate inthe target age group is 10% or higher (WHO,2002).Strengths and LimitationsOne advantage of this indicator is that it can beused to uncover reasons for dropout. For programpurposes, EPI managers will find it important toknow, for example, whether high dropout ratesoccur because mothers do not remember to bringchildren to the health center for subsequent doses,or because mothers are unaware of the need for asubsequent dose of the vaccine, or because ofImmunization151

vaccine stock-out, or some other reason. Each ofthese reasons has different implications forprogramming.This indicator can be easily calculated fromadministrative data monitoring the number ofdoses of DTP1 and DTP3 vaccine that areadministered to the target population. Estimatesbased on routine data are relatively inexpensive asthey make use of data that are currently beingcollected. The indicator also does not requireestimates of the size of the target population, amajor problem in calculating and interpretingimmunization coverage rates from routine data.Using routine data, the indicator can be calculatedat multiple levels: national, regional, district, andhealth facility. As dropout rates can be calculatedat the health facility level without requiring datafrom additional sources, they can allow a healthfacility to monitor its own performance withoutthe need for higher level monitoring andevaluation specialists. However, when routine dataare aggregated across various health facilities anddistricts, caution needs to be exercised in their usedue to potential problems with the completenessand timeliness of reporting (e.g., the proportionof sites submitting reports) and the degree towhich data from the private sector are incorporatedinto the reporting system.Population-based surveys provide morerepresentative data for estimating dropout and canbe helpful in validating the routine reportingsystem. However, conducting a survey is expensiveand not done frequently. Furthermore, care shouldbe taken when designing cluster surveys to ensuregeographical representation and avoid selectionbias. Card-based data increase the validity ofsurvey estimates, but the vaccination card may notbe widely available.This indicator cannot stand on its own and mustbe interpreted in the light of actual coverage levels.The indicator does not give a complete picture ofdropouts that may be occurring between otherantigens (USAID, 2003). Consequently, programstypically calculate the following additional dropoutrates from routine data:(1) BCG to DTP1 dropout rate:Numerator: Cumulative total of BCG minuscumulative total of DTP3 immunizationsgiven to children below the age of one yearduring a specified reference period (forexample, during the past 12 months)Denominator: Cumulative total of BCGimmunizations given to children below the ageof one year during the same periodA high BCG to DTP1 dropout rate may indicatemissed opportunities to give the vaccination cardand to provide information on where and whento bring the baby for other vaccinations. Possiblesolutions include making sure that mothers whogive birth at maternity centers and whose babiesreceive BCG are given: (a) a vaccination card forthe baby; (b) information on where and when tobring the baby for other vaccinations; and (c)encouragement to get the baby immunized againas soon as he or she reaches six weeks of age.(2) DTP3 to measles dropout rate:Numerator: Cumulative total of DTP3 minuscumulative total of measles immunizationsgiven to children below the age of one yearduring a specified reference period (forexample, during the past 12 months)Denominator: Cumulative total of DTP3immunizations given to children below the ageof one year during the same periodA high DTP3 to measles dropout rate mayindicate poor health worker and caretakercommunication. Possible solutions includeinforming parents about preventing measles bymeans of vaccination and reminding parents whenand where to bring the child in for measles vaccine.152Chapter 4

FULLULLY IMMUNIZEDCHILDHILDHILD (FIC)DefinitionProportion of children age 12-23 months whoreceived three doses of Oral Polio Vaccine (OPV),three doses of DTP, and one dose each of BacilleCalmette-Guerin (BCG) and measles vaccinesbefore age 12 months.*Numerator: Number of children age 12-23months who received three doses of OPV,three doses of DTP, and one dose each of BCGand measles vaccine before age 12 months.Denominator: Total number of children aged12-23 months survey.Measurement ToolsPopulation-based surveys (DHS, EPI ClusterSurvey, KPC, MICS)What It MeasuresThis indicator measures the success of theimmunization program in delivering allrecommended vaccines in the childhood schedulein the first year of life. It also measures publicdemand and the perceived quality of services.How to Measure ItDuring household surveys, caretakers are askedto present the child’s vaccination card to theinterviewer so that it can be used to record theactual dates at which vaccinations were given. Ifa card is available, the interviewer is required to* Note that the definition of a “fully immunized child” shouldcorrespond to the national immunization policy in a givencountry. The definition will change as new and underutilizedvaccines (such as HepB, Hib and Yellow Fever) are introduced.Immunizationcopy carefully the dates on which the child receivedvaccinations against each of the six diseasestargeted by the EPI program, namely tuberculosis,diphtheria, pertussis, tetanus, poliomyelitis andmeasles.If the mother/caretaker cannot produce avaccination card for the child, she is asked whetherthe child has received a vaccination againsttuberculosis (BCG); diphtheria, tetanus andpertussis (whooping cough) (DTP); poliomyelitis(OPV); and measles. In many surveys,interviewers also check to see whether a BCG scaris present. For OPV and measles vaccine, it isimportant to specify routine immunizationsystems, and differentiate from supplementalimmunization activities, as the source where thechild received immunization. For DTP and OPV(as well as HepB and Hib, where these have beenincorporated into the routine nationalimmunization program) information is obtainedon the number of doses given.The source of the data needs to be noted for eachchild surveyed. Card-based information increasesthe validity of the data assuming that theinformation on the card is complete and accurate.However, the absence of a card does not necessarilymean that the child in not vaccinated and it isrecommended that coverage estimation be basedon both card and history information.For a child to be fully immunized, s/he shouldreceive at least one dose of BCG, three doses ofOPV, three doses of DTP, and one dose of measlesvaccine before the first birthday. The indicatorcan be measured at the district or national level.If a dose of OPV is given at a health facility withinthe first two weeks of life (this dose being referredto as OPV0), this is not counted as one of thethree doses required to achieve full immunization.153

With new vaccines being introduced, what itmeans to be a “fully immunized child” may differby country. Thus, global comparison of thisindicator may be limited in the future.Strengths and LimitationsThis indicator is useful for measuring progresstoward the EPI goal of reducing morbidity,disability, and mortality due to six commonvaccine-preventable diseases. However, theindicator is only a measure of completion of therecommended immunization schedule, and doesnot measure protection. The impact ofimmunization on disease is dependent on thetiming and number of doses received, as well asthe efficacy of vaccine. The indicator does notreflect whether vaccines are given at therecommended ages or at the recommendedminimum interval of four weeks betweenconsecutive doses of DTP and OPV (and HEPBand Hib, if included in national definitions). Theimpact on disease may also be higher than whatFIC would lead you to expect. For example,measles coverage might be 60%, but FIC mightbe only 40% – as it is in many countries in Africabecause children have not received all the DTPdoses.With the addition of new vaccines such as HEPBto the WHO recommended immunizationschedule for children below the age of one year,the current definition of “fully immunized child”is problematic and comparisons across countriesare limited.While complete vaccination coverage before theage of 12 months is the preferred immunizationindicator, it is generally not available from routineservice statistics. Information can usually bederived from population-based surveys. Surveysare expensive and done infrequently; so thisindicator can only be estimated every 3-5 years(though in some countries, such as Bangladesh, itis estimated every two years). The absence ofvaccination cards limits the reliability of thisindicator. Care should be taken when designingcluster surveys to ensure representation and toavoid selection bias by following the protocol.154Chapter 4

VACCINEWASASTAGERATEDefinitionProportion of vaccine that is supplied but neveradministered to a child or mother during aspecified reference period (e.g. one year).Numerator: Doses supplied minus dosesadministered x 100.Denominator: Doses supplied.“Doses supplied” is calculated from stock recordsby adding the starting balance of usable vaccinedoses at the beginning of a designated period tothe new doses received during that period andsubtracting the balance remaining at the end ofthe period: Starting balance + doses received – balanceremaining = doses supplied (USAID, 2003).Measurement ToolsVaccine supply and distribution reviews; vaccinestock control ledgersWhat It MeasuresThis indicator provides a measure of the qualityand efficiency of the immunization service deliverysystem. Wastage may be due to service delivery/programmatic reasons: for example, it may bepossible to extract only 17 doses out of a 20-dosevial because some vaccine is left unused in the“dead space” of each syringe. After reconstitution,freeze-dried vaccines (like measles vaccine andBCG) must be discarded after six hours. Wastagemay occur if few children are immunized from avial, leading to the unused portions beingdiscarded. In addition, wastage may occur ifreserve vials with doses remain after ice in a vaccinecarrier has run out and there is no Vaccine VialMonitor (VVM) available. This may be a problemin very hot climates. Wastage may also be causedby logistics problems or incorrect handling ofvaccines; cold chain breakdown; freezing of DTP,ImmunizationHBV, and TT; and a manufacturer providing shortexpiry dated vaccine (which cannot possibly bedistributed and used before expiry) (WHO, 2001).How to Measure ItThis indicator is measured by reviewing vaccinesupply and distribution records and/or vaccinestock control ledgers, on a regular basis. Thenumber of doses supplied needs to be calculatedas indicated above before wastage can becomputed. As some level of wastage is expectedeven in the best system, this indicator is notexpected to ever reach zero. Remote areas mayneed to open more vials per population than urbanareas, thus programs in rural areas may need toaccept higher wastage rates to increase coverage.The acceptable wastage level depends on eachprogram and is based on experience and analysisof local situations (for example, whether a countryhas adopted the WHO policy on the continueduse of opened multi-dose vials of certain vaccineson subsequent days). In general, wastage rateshigher than 20% can be an indicator of programproblems discussed below.Strengths and LimitationsThis indicator is useful in forecasting vaccine needsand can be used by program managers to uncoverprogram problems (e.g., repeated instances oflower-than-planned attendance at immunizationsessions; poor stock management; cold chainfailure that exposes vaccines to unacceptably highor low extremes of temperature; incorrect mixingof freeze-dried vaccine; incorrect dosage [e.g., theadministration of three drops of OPV instead oftwo, or the injection of 0.6 ml of vaccine insteadof 0.5 ml]; inappropriately large vial sizes forvaccines other than BCG which is only availablein 20 dose vials; or failure to comply with a multidosevial policy, etc.).155

The vaccine wastage rate may also be used as afinancial monitoring indicator (e.g., to calculatethe cost of a fully immunized child with andwithout wastage). However, using vaccine wastagesolely as a tool for economic/financial monitoringis not useful. A poorly functioning logistics systemwill not have the information required to calculatethis indicator. Depending on the country,determining vaccine wastage for diphtheria,tetanus, and pertussis may be difficult because theyare administered through a number of differentvaccine combinations.156Chapter 4

RefererencesencesBonhoeffer J, Kohl K, Chen R, Duclos P, Heijbel H, Heininger U, Jefferson T, Loupi E, and TheBrighton Collaboration. (2002). The Brighton Collaboration: Addressing the Need for StandardizedCase Definitions of Adverse Events Following Immunization (AEFI). Vaccine, 21:298-302.Bos E, and Batson A. (2000). Using Immunization Coverage Rates for Monitoring Health SectorPerformance: Measurement and Interpretation Issues. Washington, DC: The World Bank. HNPDiscussion Paper.CDC. (1999). Global Measles Control and Regional Elimination, 1998-1999. Morbidity and MortalityWeekly Report, 48/49:1124-1130.GAVI. (2001). Global Alliance for Vaccines and Immunization. Geneva, Switzerland: World HealthOrganization. Fact Sheet No. 169.Goldman N, and Pebley AR. (1994). Health Cards, Maternal Reports and the Measurement ofImmunization Coverage: The Example of Guatemala. Social Science and Medicine, 38(8):1075-89.Hauri AM, Armstrong GL, and Hutin YJF. (2003). In: Comparative Quantifications of Health Risks:Global and Regional Burden of Diseases Attributable to Selected Major Risk Factors. ContaminatedInjections in Health Care Settings. Geneva, Switzerland: World Health Organization.Kane A, Lloyd J, Zaffran M, Simonsen L, & Kane M. (1999). Transmission of Hepatitis B, Hepatitis Cand Human Immunodeficiency Viruses Through Unsafe Injections in the Developing World: Model-basedRegional Estimates. Geneva, Switzerland: World Health Organization. Bulletin of the World HealthOrganization, 77(10):801-807.Nelson KE, Williams CM, and Graham NMH. (2001). Infectious Disease Epidemiology – Theory andPractice. Gaithersburg, MD: Aspen Publication.Simonsen L, Kane A, Lloyd J, Zaffran M, and Kane M. (1999). Unsafe Injections in the DevelopingWorld and Transmission of Bloodborne Pathogens: A Review. Geneva, Switzerland: World HealthOrganization. Bulletin of the World Health Organization, 77(10):789-800.Szklo M, and Nieto FJ. (2000). Epidemiology – Beyond the Basics. Gaithersburg, MD: Aspen Publication.USAID. (2003). Immunization Essentials: A Practical Field Guide. Washington, DC: USAID.Immunization157

WHO. (1994). Expanded Programme on Immunization – Accelerated Measles Strategies. WeeklyEpidemiology Record, 69:229-34.WHO. (1996). Using Surveillance Data and Outbreak Investigations to Strengthen Measles ImmunizationProgrammes. Retrieved from http://www.who.int/vaccines-documentsWHO. (1998). Module 11: Monitoring Immunization Coverage. Geneva, Switzerland: World HealthOrganization, Global Programme for Vaccines and Immunization, Expanded Programme onImmunization. WHO/EPI/TRAM/98.11.WHO. (1999). Indicators to Monitor Maternal Health Goals: Report of a Technical Working Group, Geneva,8-12 November 1993. Geneva, Switzerland: World Health Organization, Division of Family Health.WHO. (2000). Global Eradication of Poliomyelitis. Retrieved June 27, 2003, from http://www.who.int/vaccines-documents/WHO. (2000c). Hib Immunization: Introducing Haemophilus Influenza Type B Conjugate Vaccine intoNational Immunization Programmes. Geneva: World Health Organization, Department of Vaccinesand Biologicals. AIDE-MEMOIRE.WHO. (2000d). Maternal and Neonatal Tetanus Elimination by 2005: Strategies for Achieving andMaintaining Elimination. UNFPA/UNICEF/WHO. WHO/V&B/02.09. Retrieved from http://www.who.int/vaccines-documents/DocsPDF02/www692.pdfWHO. (2001). Hepatitis B Immunization: Introducing Hepatitis B Vaccine into National ImmunizationProgrammes. Geneva, Switzerland: World Health Organization, Department of Vaccines and Biologicals.AIDE-MEMOIRE.WHO. (2001a). Vaccine Wastage and Monitoring the EPI. Retrieved from http://www.who.int/vaccinesaccess/vacman/technet21/TC21_schnur.pdfWHO. (2001b). Tool for the Assessment of Injection Safety. Geneva, Switzerland: World HealthOrganization, Department of Vaccines and Biologicals.WHO. (2002). User’s Handbook for Vaccine Cold Rooms and Freezer Rooms. Geneva, Switzerland:World Health Organization. WHO/V&B/02.31.WHO. (2002a). The Immunization Data Quality Audit (DQA) Procedure. Geneva, Switzerland: WorldHealth Organization158Chapter 4

WHO. (2002b). Increasing Immunization Coverage at the Health Facility Level. Geneva, Switzerland:World Health Organization, Department of Vaccines and Biologicals.WHO. (2003). Hepatitis B Vaccine. Retrieved June 16, 2003, from http://www.who.int/vaccines/en/hepatitisb.shtmlWHO. (2003b). Current Epidemiological Situation of Reported Epidemic Prone Diseases in the Region(Year 2003). Retrieved from http://www.afro.who.int/csr/epr/epidemic_prone_diseases/weekly_feedback1-7_2003.pdfChapter 4. Immunization159

5INTEGRTEGRATETEDDISEASESURURVEVEILILLANCEAND RESPONSEIndicators:• Proportion of health facilities submitting weekly/monthly surveillancereports on time to the district level• Proportion of districts submitting weekly/monthly surveillance reportson time to the next level• Proportion of cases of diseases selected for case-based surveillance whichwere reported to the district using case-based or line listing forms• Proportion of suspected outbreaks of epidemic-prone diseases notified tothe next level within two days of surpassing the epidemic threshold• Proportion of districts with current trend analysis (line graphs) for selectedpriority diseases• Proportion of reports of investigated outbreaks that include case-baseddata recorded and analyzed• Proportion of outbreaks of epidemic-prone diseases with laboratory results• Proportion of confirmed outbreaks with recommended response• Attack rate• Case fatality rate for outbreaks of epidemic-prone diseases

HAPTER 5. ICHAPTER5. INTEGRANTEGRATEDTED DISEASESURURVEILLVEILLANANCEAND RESPONSESurveillance is the ongoing process of systematiccollection, collation, analysis, andinterpretation of data with prompt disseminationto those who need to know for relevant action tobe taken. A well functioning disease surveillancesystem is essential for the control of communicablediseases, the most common causes of death anddisability in some developing regions of the world,notably Africa. Accurate surveillance data providecontinuous information about disease trends overtime and may serve to identify outbreaks in theearly stages of development. If surveillance dataare linked to laboratory results, feedback fromthese activities can provide public health officialswith an understanding of how an outbreak isprogressing and facilitate the design of effectivedisease control strategies. Adequate measlessurveillance, for example, permits furtherevaluation of immunization coverage as well asthe implementation of appropriate measures toimprove disease control, such as house-to-houseimmunization for high-risk populations, preoutbreakacceleration of activities, and masscampaigns to interrupt measles transmission andeliminate the disease.The core activities for an effective surveillancesystem are the following:• Detection (identifying cases and outbreaks)• Registration• Confirmation (epidemiological and laboratoryconfirmation)• Reporting (early warning and routine)• Analysis and interpretation (preparing andperiodically updating graphs, tables, and chartsto describe time, person, and place for reporteddiseases and conditions; identifying unusualtrends or patterns or the exceeding of athreshold value; interpreting results; discussingpossible public health action)• Responseo Control/response (case management,contact tracing, infection controlmeasures, immunization activities,improvement of preventive and controlmeasures [vector and/or environmentalcontrol], community information, andeducating or alerting nearby areas anddistricts)o Outbreak investigation (case findings[records, active surveillance], collectionand transport of specimens, confirmatorytesting, and interpretation of results[epidemiological and laboratory])o Program adjustmento Changes in policy and planning• Feedback• Evaluation and monitoringThese activities are made possible by a number ofsupport functions that lead to better performanceof the core surveillance activities:• Setting standards (e.g., case definitions,standard case management guidelines,standard procedures for investigation)• Training (surveillance, epidemiology,laboratory)• Supervision• Communications systems (e.g., radio, fax, e-mail, phone, health updates)• Providing resources (human – appropriatenumber with adequate skills andcompetencies; material – vehicles, laboratoryequipment, supplies, etc.; financial)An assessment of the capacity to carry out coreand support functions of surveillance and responsefor each priority disease or group of diseases at allChapter 5. Integrated Disease Surveillance and Response161

levels of the health system (central, regional/provincial, district or equivalent, health facility) isan integral component of any assessment ofnational communicable disease surveillance andresponse. The following facets of the systemshould also be addressed in order to determinewhether the system is meeting its objectives:• The public health importance of the healthevent(s) under surveillance• The objectives and operation of the system• The system’s usefulness• Attributes of the system• Cost or resource requirements of the systemAs specified by the CDC (2001), importantattributes include:• Simplicity – the ease of operation of thesystem as a whole and each of its components• Flexibility – ability to accommodate changesin operating conditions and information needs• Data quality – completeness and validity ofthe data collected and recorded• Acceptability – willingness of individuals andorganizations to participate in the system• Sensitivity – ability to detect the cases orhealth events or outbreaks it is intended todetect• Predictive Value Positive (PVP) – mostlyaffected by the system’s specificity, PVP is theproportion of reported cases (or outbreaks)which truly are cases (or outbreaks)• Representativeness – extent to which thesystem accurately portrays the incidence ofhealth events in a population by time, place,and person• Timeliness – availability of data in time forappropriate action• Stability – reliability and availability of thesystem (operates properly without failure)Assuming that all levels of the health system areinvolved in conducting surveillance activities fordetecting and responding to priority diseases andconditions, the information flow in an integrateddisease surveillance system is depicted in Figure5.1 on the following page.Methodological ChallengesFollowing are key methodological challenges andissues in monitoring and evaluating the corefunctions of disease surveillance systems:Non-standard Case DefinitionsA case definition is a standard criteria used todecide if a person has a particular disease or if thecase can be considered for reporting andinvestigation. In investigating an outbreak of apriority disease, the specificity of the casedefinition needs to be high to have an accuratecount of those who have the disease. Componentsof the case definition may include informationabout exposure (e.g., time and place), laboratoryfindings, and clinical symptoms. In routinesurveillance as well as outbreak investigations,cases of disease are commonly separated into“confirmed cases” and “probable cases.” Confirmedcases are generally those that have been confirmedby laboratory results, and probable cases are thosethat have certain symptoms meeting a clinical casedefinition but have not been confirmed bylaboratory tests. When all health workers usestandard case definitions for reporting cases ofpriority diseases, there is consistency in reportingand it is easier for public health officials to followtrends in disease and to recognize epidemics. Inaddition, case definitions may have differentsensitivity and specificity, limiting meaningfulcomparative analyses. WHO has providedstandardized case definitions for the differentpriority diseases and it is important to ensure thathealth workers are encouraged to use them (seeAnnex 5.1 on page 195).162Chapter 5

Figure 5.1. Information flow in an integrated disease surveillance system (IDS).“An ill person presents to medical attention. Information about the patient is recorded in a register.The register is updated daily to include information for both inpatients and outpatients. At aminimum, the following data are collected: the patient's ID number, date of onset, date of presentationat the facility, date of discharge (inpatient only), village (location), age, gender, diagnosis, treatment,and outcome (inpatient only).If the clinician suspects a disease or condition that is targeted for elimination or eradication, or if thedisease has high epidemic potential, the disease is reported immediately to the designated healthstaff in the health facility and at the district level. The health facility should begin a response to thesuspected outbreak. At the same time, the district takes steps to investigate and confirm the outbreak.The investigation results are used to plan a response action with the health facility.Periodically, once a month, weekly, quarterly, or annually, the health facility summarizes the numberof cases and deaths for each routinely reported IDS condition and reports the totals to the district.The health facility performs some analysis of the data such as keeping trend lines for selectedpriority diseases or conditions and observing whether certain thresholds are passed to alert staff totake action. One action that is taken if an outbreak is suspected is to obtain laboratory confirmation.Laboratory specimens are obtained and the following data are documented: type of specimen, dateobtained, date sent to the lab, condition of specimen when received in the lab (good or poor), andlab results.At the district level, data are compiled monthly for each of the IDS conditions. The district preparesanalyses of time, place, and characteristics of the patients such as age and gender for both outpatientsand inpatients. The results are sent to either the regional level or the central level.The district uses the data to plot graphically the routine surveillance trends and epidemic curves forIDS conditions. In addition, the district maintains a log of suspected outbreaks reported by healthfacilities. This list documents the nature of the potential outbreak, the number of possible cases, thedates of investigations and actions taken by the district. It also includes any findings of investigationsled by the district, regional, or national levels.The district surveillance focal point provides disease-specific data and information to each diseaseprevention program.”(WHO-AFRO and CDC, 2001, 5)Accuracy, Completeness, and Timeliness ofReportingIn active surveillance systems, the followingassumptions are often made:(1) Cases are occurring in the community.(2) Persons who are cases seek medical attentionor otherwise come to the attention ofinstitutions subject to reporting requirements.(3) The condition is recognized by the provideror health facility.(4) Cases are not reported because filling outreporting forms or notifying to the next higherlevel is too much trouble.(5) If the administrative reporting burden forhealth facilities is reduced, cases will bereported.Integrated Disease Surveillance and Response163

Many of these assumptions are rarely met indeveloping countries, with implications for theaccuracy, completeness, and timeliness ofreporting. In areas characterized by poor accessto and utilization of health services, if thecommunity does not know how to notify healthauthorities when priority diseases or unusualhealth events occur, suspected cases will not beseen at the health facility, and cases will not bereported. It is important, therefore, for communityhealth workers, traditional healers, birthattendants, and community leaders to know howto recognize and report selected priority diseasesor conditions to the health facility for treatment.Timely notification is essential so that publichealth action can be taken to limit the spread ofthe disease. Methods for dealing with theselimitations include distributing simplified casedefinitions for use in community surveillance (seeWHO-AFRO and CDC, 2001, p. 32-33 for anexample) and responding effectively to communityreports in order to encourage communityparticipation in the disease surveillance system.Problems of completeness of reporting may alsooccur if some districts do not send reports whenno new cases of reportable diseases are found in aparticular reporting period. The reporting of “zerocases” when no cases have been detected by thereporting unit allows the next level of the reportingsystem to be sure that data sent by participatingunits have not been lost or that the participatingunit has not forgotten to report. If zero-reportingis not accompanied by efforts at case finding, itwould be difficult to know whether the absenceof reported cases means the absence of disease inthe population.RepresentationAnother dimension of completeness, which hasimplications for the validity of surveillance data,relates to the issue of representation (that is,whether the probability of reporting a disease isthe same within subgroups of the population orin different populations). Issues of validity canoccur if some parts of a country are not coveredby the surveillance system effectively. In addition,the quality of case ascertainment can be affectedby the differential reporting of cases in associationwith different characteristics of the person – inthe sense that cases among certain subpopulationsmay be less likely to be reported than among othergroups (Romaguera, German, and Klaucke, 2000).For example, cases in adults or children who arenot legal residents of a country or who are refugeesare likely to be underreported or inadequatelyinvestigated due to a number of factors. Illegalresidents may not have access to medical care, maybe transient or may avoid contact with authoritiesif they do not possess proper documentation.Changes in reporting practices over time can alsointroduce bias into the system, making it difficultto follow long-term trends or establish baselinerates to be used for the recognition of outbreaks.The ability to detect imported cases is an indirectmeasure of the quality of case ascertainment atthe national level. At the district level, noimportations might occur and the absence ofreported cases may reflect either the absence ofdisease or the absence of efforts to identify cases.In nations that border countries where a particulardisease is endemic and where there is substantialcross-border movement, failure to detect importedcases of the priority disease may suggest that thenational level surveillance system is not sensitiveenough to detect individual cases.Errors in Descriptive Information about aReported CaseEven if a case of a notifiable disease has beenidentified and reported, there may be errors in thecollection and recording of descriptive informationabout the case. These errors can be introduced atany stage of the reporting and assessment process.Information commonly collected by surveillancesystems includes the demographic characteristicsof the affected individual, details about the healthevent, and the presence or absence of definedpotential risk factors. The quality of data is alsoinfluenced by the simplicity of the reporting forms,the level of training and supervision of persons164Chapter 5

who complete the surveillance forms, and the careexercised in data management (Romaguera,German and Klaucke, 2000).Lack of Laboratory Support for OutbreakConfirmation or Patient ManagementIn low-income settings, laboratory services maybe unavailable or nonstandardized. Logistics ofgetting specimens to the laboratory may also bedifficult. Strategies used by the polio eradicationprogram to overcome this problem include usingthe private sector to help transport specimens tothe lab and using the laboratory results to ensurethat surveillance systems results are supervised andaccurate (White and McDonnel, 2000). However,expanding laboratory improvements built fromvertical programs associated with polio eradicationand neonatal tetanus control within an integrateddisease surveillance and response system remainsa significant challenge.Lack of External StandardsThe rate of Acute Flaccid Paralysis (AFP) amongchildren less than 15 years of age is a powerfulindicator of the adequacy of polio surveillance.Unfortunately, similar external standards do notexist for many other vaccine preventable diseases.For diseases such as Hib, which occurs amongchildren under five years of age as well as adults,it may be necessary for child health programs tomonitor the absence of Hib cases in any age groupsto assess whether the surveillance system isadequate (Wharton and Ching, 2002).Infrastructure and Communication ConstraintsWhite and McDonnel (2000) have described indetail infrastructure and communicationconstraints that may affect the quality ofsurveillance data in low- and middle-incomecountries. These constraints may includeidentifying personnel to conduct surveillance andensuring transportation and operating expenses forhealth staff to investigate cases, trace contacts, ortransport specimens to laboratories are importantconstraints. In some situations, health facilitiesmay not have sufficient reporting forms and healthpersonnel may not be adequately trained to usethese forms to report priority diseases within thesurveillance system (White and McDonnel, 2000).A limited resource base and the lack of a regularsupply of electricity may not permit the creationof a functional computer-based system forreporting. Where there is reliance on a handwrittenreport system, it may be more difficult tomaintain quality control during the collection andtabulation of data or stimulate review and use ofinformation for decision-making. Over time, theincreased availability of computers in low-incomecountries at the national, regional, and districtlevels has permitted greater electronic reporting.Selection of IndicatorsOver the years, many of these vaccine-preventabledisease control programs have developed their ownsurveillance standards and systems to obtain timelyand reliable information for monitoring trends indisease occurrence, predicting or providing earlydetection of outbreaks, and initiating appropriatepublic health action and response (see Annex 5.2).Specialized surveillance systems (e.g., AFP, HIV/AIDS) are important, especially when surveillancemethods are complex and the systems have specificinformation needs. However, all surveillancesystems involve the same universal functions andcommon support functions. With the coexistenceof various surveillance activities funded andmanaged by different control programs, which aresometimes based in different institutions, theoverall surveillance functions at the national levelcan become inefficient. In such cases, healthpersonnel participate in multiple systems, usedifferent surveillance methods, terminology,reporting forms and frequency, resulting in extracosts and often work overload for health staff(WHO, 2001). An integrated disease surveillancesystem allows all surveillance activities, whetherfor acute flaccid paralysis, measles, neonataltetanus, or other diseases, to be coordinated andstreamlined and takes advantage of similarsurveillance functions, skills, resources, and targetpopulations (CDC, 2003).Integrated Disease Surveillance and Response165

The World Health Organization (WHO)Regional Office for Africa (AFRO) has adoptedan integrated disease surveillance and response(IDSR) strategy linking community, health facility,districts and national levels. The objectives ofintegrated disease surveillance are to provide arational basis for decision-making andimplementing public health interventions that areefficacious in responding to priority communicablediseases. The IDSR was adopted by member statesof the African Region in 1998. WHO suggests19 communicable diseases and conditions forintegrated diseases surveillance in the Africanregion (WHO/CDC, 2001). These include (1)epidemic-prone diseases (cholera, diarrhea withblood [shigella], measles, meningitis, plague, viralhemorrhagic fevers, and yellow fever); (2) diseasestargeted for eradication and elimination (acuteflaccid paralysis [AFP]/polio, dracunculiasis,leprosy, neonatal tetanus); (3) other diseases of publichealth importance (pneumonia in children less thanfive years of age, new AIDS cases, malaria,onchocerciasis, sexually transmitted infections[STIs], trypanosomiasis, and tuberculosis).In February 2001, the IDSR task force formed ajoint WHO and CDC working group to developand test indicators for monitoring and evaluatingIDSR, making it possible to assess the surveillancesystem as a whole and approach systemdevelopment and strengthening in a coordinatedway (WHO, 2001). The working group proposeda list of core indicators for testing at the IDSRtask force meeting in May 2001 (CDC, 2003).From January to June 2002, the working groupcollaborated with Ministries of Health in Ugandaand Mozambique to conduct pre-tests of theindicators at national, provincial, and district levels.This section of the guide outlines the core indicatorsdeveloped by WHO-AFRO and CDC formonitoring the implementation of IDSR in theAfrica region. The indicators are presented below:• Proportion of health facilities submittingweekly/monthly surveillance reports on timeto the district level• Proportion of districts submitting weekly/monthly surveillance reports on time to thenext level• Proportion of cases of diseases selected forcase-based surveillance which were reportedto the district using case-based or line listingforms• Proportion of suspected outbreaks ofepidemic-prone diseases notified to the nextlevel within two days of surpassing theepidemic threshold• Proportion of districts with current trendanalysis (line graphs) for selected prioritydiseases• Proportion of reports of investigated outbreaksthat include case-based data recorded andanalyzed• Proportion of outbreaks of epidemic-pronediseases with laboratory results• Proportion of confirmed outbreaks withrecommended response• Attack rate• Case fatality rate for outbreaks of epidemicpronediseasesThough initially developed for the Africa region,the IDSR indicators presented in this section canbe adapted to national surveillance needs and acountry’s own disease control priorities, objectives,and strategies. The indicators can help identifygaps and opportunities in performing the corefunctions of surveillance. They can also helpdetermine country needs as regards strengtheningthe surveillance system for communicable diseasecontrol and prevention, and provide baselineinformation against which to measure progress instrengthening integrated disease surveillance andresponse systems.It must be noted that not all diseasesrecommended for integrated disease surveillanceare under-five priority diseases and not all vaccine-166Chapter 5

preventable diseases of the EPI are part of theIDSR strategy. Diphtheria, for example, is awidespread severe infectious disease that has thepotential for epidemics and primary preventionof the disease is by ensuring high populationimmunity through immunization. Recentepidemics have highlighted the need for adequatesurveillance and epidemic preparedness (WHO,1998). Pertussis or whooping cough is also a majorcause of childhood morbidity or mortality in thedeveloping world. An estimated 20-40 millioncases, 90% of which occur in developing countries,and 200,000 to 300,000 deaths occur annually.Although pertussis may occur at any age, mostcases of serious disease and a majority of fatalitiesoccur in early infancy. Case fatality rates indeveloping countries are estimated to be as highas 4% in infants (WHO, 2001). It is vital,nonetheless, to monitor the implementation ofsurveillance activities for these diseases within theframework of integration.Integrated Disease Surveillance and Response167

PROPOROPORTIONOF HEALEALTHFACILITIESSUBMITTINUBMITTING WEEKLEEKLY/MONTHLONTHLY SURURVEILLVEILLANANCEREPOREPORTS ON TIMETO THEDISISTRICTLEVELCore e IDSR Qualituality y IndicatorDefinitionProportion of health facilities submitting weekly/monthly surveillance reports on time to the districtduring a specified period.Numerator: Number of health facilitiessubmitting weekly/monthly surveillance reportson time to the district during a specified period.Denominator: Total number of health facilitiesexpected to submit weekly/monthly surveillancereports to the district in the same period.“On time” means that the forms are received withina specified time from the end of the reportingperiod. Common reporting periods are “within 7days after the start of a new month,” “within 14days of the start of a new quarter,” or other setperiod. National policy determines whether thedata from health facilities and districts are reportedimmediately, weekly, monthly, or quarterly. Forepidemic-prone diseases, notification should beimmediate.Measurement ToolsAdministrative records (such as a table forrecording timeliness and completeness of monthlyreporting from the health facility to the district[see Annex 5.3 on page 199 for a sample form forrecording the timeliness and completeness ofmonthly reporting from the health facility to thedistrict]); computerized HIS/MIS databasesWhat It MeasuresThis indicator provides a measure of the extent towhich health facilities make disease surveillancedata accessible to the district in a timely fashion.It is an important measure of the quality of thereporting system at the health facility level. Theexistence of timely information is a preconditionfor efficient response to any public health problemand should therefore be monitored. The qualifier“on time” in the definition of the indicatorhighlights the need to monitor cases and deathsfor each routinely reported IDS condition at a paceconsistent with the time frame that districtmanagers have set for planning, monitoring,evaluation, and decision-making.How to Measure ItMeasurement of this indicator requires recordsindicating exact dates when reports are sent fromhealth facilities and received at the district healthoffice where they are aggregated. Periodically(once a month, quarterly, or annually), the healthfacility summarizes the number of cases and deathsfor each routinely reported IDS condition andreports the totals to the district. At the districtlevel, the dates on which reports are received areroutinely recorded and reviewed during theanalysis of routine and case-based data. Therecords of reports that have been received are usedto:• Measure how many reporting units (i.e., healthfacilities) submitted reports for a given month(completeness of reporting units);• Identify which reporting units have reported;and• Measure how many reports were submittedon time.This indicator can be measured every three monthsby a district supervisor at the health facility level.A rate of timeliness of 80% or more is regarded ashighly satisfactory. When the surveillance systemis good, the rate of timeliness should approachIntegrated Disease Surveillance and Response169

100%. Not receiving an adequate performance onthis indicator is a trigger for problem solving.Countries may choose to adapt the target andmake it incremental as performance problems areaddressed.Strengths and LimitationsMeasurement of this indicator serves a number ofprogrammatic purposes. If data indicate that ahealth facility has not provided a report, or if thereport is not on time, the surveillance focus pointat the facility should be contacted to work withthe designated staff to identify the cause of theproblem and develop solutions. These solutionsmay include providing resources (including asufficient and reliable supply of forms for reportingthe required information, on-the-job training tostaff at the facility regarding reporting procedures,etc.). However, deadlines for submitting routinereports must be reasonable given the particularchallenges to delivery in many developing countrysettings.The indicator assumes that IDS reporting formsare available for reporting at the health facilitylevel. It also assumes that standard case definitionsspecified by national policy have been widelydisseminated and that health facility staff knowhow to use them to detect priority diseases. Itshould be noted that this indicator does not assessthe quality of the reports themselves or theirrepresentativeness. This is crucial for evaluatingthe reporting system information. Furthermore,the indicator does not look at the accuracy orcompleteness of reporting. For a surveillancesystem to be considered good quality, reportingshould be not only timely but also complete.Depending on the comprehensiveness of thesurveillance system, the indicator may or may notinclude reporting from private health facilities ornongovernmental organizations.170Chapter 5

PROPOROPORTIONOF DISISTRICTTRICTS SUBMITTINUBMITTING WEEKLEEKLY/MONTHLONTHLY SURURVEILLVEILLANANCEREPOREPORTS ON TIMETOTHE NEXTLEVELCore e IDSR Qualituality y IndicatorDefinitefinitionionProportion of districts submitting weekly/monthlysurveillance reports on time to the next level(provincial/regional) during a specified period.Numerator: Number of districts submittingweekly/monthly surveillance reports on timeto the next level (provincial/regional) duringa specified period.Denominator: Total number of districtsexpected to submit weekly/monthlysurveillance reports to the next level(provincial/regional) in the same period.“On time” means that the forms are received withina specified time from the end of the reportingperiod. Common reporting periods are “within 7days after the start of a new month,” “within 14days of the start of a new quarter,” or other setperiod. National policy determines whether thedata from districts and health facilities are reportedimmediately, weekly, monthly, or quarterly. Forepidemic reporting, notification should beimmediate.Measurement ToolsAdministrative records (such as a form forrecording timeliness and completeness of monthlyreporting from the district to the provincial orregional level); computerized HIS/MIS databasesWhat It MeasuresThis indicator provides a measure of the extent towhich district health offices submit reports/summary data to the next level in a timely fashion.It is an important indicator of the quality of thereporting system. The existence of timelyinformation is a precondition for the use ofinformation for monitoring and response. Thequalifier “on time” in the definition of the indicatorhighlights the need to monitor cases and deathsfor each routinely reported IDS condition at a paceconsistent with the time frame that provincial/regional managers have set for planning,monitoring, evaluation, and decision-making.How to Measure ItMeasurement of this indicator involvesmonitoring the receipt of reports from districthealth offices to evaluate the timeliness ofreporting to the next level. A monitoring toolsuch as a record of reports received may be usedto monitor timeliness of reporting from districts.This indicator can be measured every three monthsby a provincial or regional supervisor at the districtlevel. A proportion of 80% or more is regarded ashighly satisfactory in terms of timeliness ofreporting from the district to the next level. Whenthe surveillance system is good, the rate fortimeliness should approach 100%. Not receivingan adequate performance on this indicator is atrigger for problem solving. Countries may chooseto adapt the target and make it incremental asperformance problems are addressed.Strengths and LimitationsAssessment of the timeliness of reporting for eachprovince/region enables the identification ofconstituent districts that scored low on thisindicator. Factors associated with delays inreporting may be determined by a routinesupervisory visit to district health offices andconcentrated efforts made to strengthen thereporting system through training and carefulplanning.Integrated Disease Surveillance and Response171

It should be noted that this indicator does notprovide a measure of the completeness of reportingor the geographic representation of the reports.For a surveillance system to be considered a goodquality, reporting should be not only timely butalso complete. Furthermore, the indicator doesnot measure the extent to which provinces/regionsprocess the reported data in a timely manner. Theindicator assumes that different IDS reportingforms are available at the district level. It alsoassumes that standard case definitions specifiedby national policy have been widely disseminatedand that health facility and district-level staff knowhow to use them to detect priority diseases.Depending on the comprehensiveness of thesurveillance system, the indicator may or may notinclude reporting from private health facilities ornongovernmental organizations172Chapter 5

PROPOROPORTIONOF CASESOF DISEASESSELECTEDFORCASEASE-BASEDSURURVEILLVEILLANANCEWHICHICH WEREREPOREPORTEDTO THEDISISTRICTUSINSING CASEASE-B-BASEDOR LINELISISTINTING FORMS-BCore e IDSR Qualituality y IndicatorDefinitionProportion of cases of diseases selected for casebasedsurveillance which were reported to thedistrict using case-based or line listing forms.Numerator: Number of cases of diseasesselected for case-based surveillance whichwere reported to the district using case-basedor line listing forms during a specified period.Denominator: Total number of cases ofdiseases selected for case-based surveillancethat occurred in health facilities in the districtin the same period.“Case:” A person who has the particular disease,health disorder, or condition which meets the casedefinitions for surveillance and outbreakinvestigation purposes.“Diseases selected for case-based surveillance:”WHO-AFRO IDSR has selected the followingdiseases for case-based surveillance: plague,measles, poliomyelitis, dracunculiasis, leprosy,neonatal tetanus, viral hemorrhagic fever, andyellow fever. The list of priority diseases selectedfor case-based surveillance could vary fromcountry to country depending on the localepidemiological situation and could include otherdiseases.“Case-based (surveillance reporting) form:” WHOrecommends a generic case-based reporting formthat can be used to report written informationabout individual cases of priority diseasesrecommended for case-based surveillance. Thetop half of the form records information aboutthe individual case and the bottom half is alaboratory transmittal slip (see Annex 5.4). Thecase-based surveillance reporting form may beadapted at the country-level. Note that somediseases (e.g., neonatal tetanus and AFP) havetheir own more detailed case investigation forms.“Line listing:” Line listing is a tool used duringepidemiological investigations to allowinvestigators to record case information and toreview and follow-up case reports or conductanalysis. Line lists are typically used instead ofindividual case-based forms to report and recordcases when several cases occur during a shortperiod (for example if more than 5 to 10 casesoccur in a week). They are also used to reportsummary totals of cases and deaths each weekwhen a large number of cases occur in a singlesuspected outbreak.Measurement ToolsAdministrative records (case-based reportingforms); computerized HIS/MIS databases; clinicregisters or patient charts; line listing forms (seeAnnex 5.5 on page 202 for a generic line listingform)What It MeasuresThis indicator measures the extent to which casebasedinformation is reported for priority diseases,which facilitates more detailed case investigation.The use of case-based information is essential forcase investigation in order to identify failure inprevention programs, causal agents, the source ofinfection, disease transmission patterns, and riskfactors related age, sex, time, place, immunizationstatus, and so-forth. Ensuring the use of standardmethods for reporting priority diseases throughoutthe system is important so that program managers,surveillance officers, and other health care staffIntegrated Disease Surveillance and Response173

can use the information to identify problems andplan appropriate responses, take action in a timelyway, and monitor disease trends in an area. Theuse of standard case-based reporting or line listingforms also allows data to be reported efficiently.How to Measure ItMeasurement of this indicator requires a reviewof surveillance reports. If the disease is one thatrequires immediate reporting or if an outbreak ofany priority disease is suspected, a case-basedreporting or line listing form should be completedat the health facility and sent to the district afterthe initial verbal report is made. The case-basedform should include the following information:• Patient’s name. If neonatal tetanus is reported,also record the name of the mother• Patient’s date of birth, if known, or age of thepatient• Patient’s locating information (address, village,neighborhood)• How to contact the patient or the parents ofthe patient if more information is needed• Patient’s gender• The date the patient was seen at the healthfacility and the date the case was reported tothe district• Date of onset of the disease• Patient’s immunization history for suspectedcases of vaccine-preventable diseases (and alsofor the mother if neonatal tetanus is suspected)• Patient’s status at the time of the report (if aninpatient, report final outcome as living ordeceased)• Date of the reportThis indicator can be measured every three monthsby a district supervisor at the health facility level.If the utilization of standard case-based reportingforms is high, this indicator should approach100%. Not receiving an adequate performance onthis indicator is a trigger for problem solving.Countries may choose to adapt the target andmake it incremental as performance problems areaddressed.Strengths and LimitationsPrecise measurement of this indicator relies onhealth staff knowing and using the standard casedefinitions as recommended by national policy.The indicator does not assess whether therecommended minimum data elements for casebaseddata are present in reports. The usefulnessof case-based data depends on whether essentialinformation is recorded on the forms. For example,if data were missing on place of occurrence, itwould be difficult to determine where cases areoccurring to identify high-risk areas or locationsof populations at risk of the disease. The indicatoralso assumes that recommended case-basedsurveillance reporting forms are available at thehealth facility level.It should be noted that this indicator does notprovide a measure of the timeliness of reporting.WHO recommends that a verbal or writtennotification of immediately reportable diseases orunusual events should reach the district within 24hours from when the case was first seen by thehealth facility. Depending on the comprehensivenessof the surveillance system, theindicator may or may not include reporting fromprivate health facilities or nongovernmentalorganizations.The indicators can be modified to reflect underfivepriority diseases as follows:• Proportion of cases of AFP which werereported to the district using case-based or linelisting forms during a specified period• Proportion of cases of measles which werereported to the district using case-based or linelisting forms during a specified period• Proportion of cases of neonatal tetanus whichwere reported to the district using case-basedor line listing forms during a specified period174Chapter 5

PROPOROPORTIONOF SUSPECTEDOUTBREAKUTBREAKS OF EPIDEMICPIDEMIC-PRONEDISEASESNOTIFIEDTO THE NEXTLEVELWITHINTWO DAYS OF SURPURPASSINASSING THE EPIDEMICTHRESHOLDCore e IDSR Qualituality y IndicatorDefinitionProportion of suspected outbreaks of epidemic-pronediseases notified to the next level (provincial/regional/national) within two days of surpassing the epidemicthreshold.Numerator: Number of suspected outbreaks ofepidemic-prone diseases notified to the next level(provincial/regional/national) within two days ofsurpassing the epidemic threshold.Denominator: Total number of suspectedoutbreaks in the district.“Suspected outbreak:” For epidemic-prone diseases,and for diseases targeted for elimination oreradication, a single case is a suspected outbreak andrequires immediate reporting.*“Epidemic-prone diseases:” These include cholera,diarrhea with blood (shigella), measles, meningitis,plague, viral hemorrhagic fever, and yellow fever.“Epidemic threshold:” An epidemic threshold is thenumber or density of susceptibles required for anepidemic to occur. For meningitis, for example, theepidemic threshold for the African belt area is anincidence of 15 cases per 100,000 inhabitants perweek over a period of two consecutive weeks.Epidemic thresholds recommended by WHO maybe adapted to meet national policies, priorities, andcapacities to respond. Note that the threshold forsome diseases will not change between healthfacilities or districts because they are thresholds forimmediately notifiable diseases and are set by nationalpolicy.Epidemic thresholds for notifying the district aboutother diseases such as shigella, malaria, measles innon-elimination countries, diarrhea with some orsevere dehydration in children less than five yearsand meningitis may be set at the health facility levelas described in Figure 5.2 on the following page.Measurement ToolsAdministrative records (e.g., district log ofsuspected outbreaks and rumors); computerizedHIS/MIS databasesWhat It MeasuresThis indicator measures the extent to whichreported outbreaks or cases are investigated,tracked, and reported to the province within thetime frame specified (i.e., within two days ofsurpassing the epidemic threshold). This isimportant to ensure that timely decisions are madewith regard to the outbreak investigation andresponse process.How to Measure ItThis indicator is measured by reviewing districtlogs of suspected outbreaks and rumors by time,place, and person which are used to record verbalor written information from health facilities orcommunities about suspected outbreaks, rumors,or unexplained events. Two sets of informationare required to calculate this indicator: (1) thedate the province was notified by the district, and(2) the date the district was notified by healthfacilities and/or communities about suspectedoutbreaks, rumors, or unexplained events. Thelogbook can be easily reviewed during asupervisory visit or when the provincial/regional/national response team wants to have informationabout how to respond to health events in thedistrict.* This is not the case for measles in the Africa region.Integrated Disease Surveillance and Response175

Figure 5.2. Steps for establishing health facility thresholds.Step 1:Step 2:Step 3:Step 4:If data from previous years are available, review trends in cases and deaths due tothese diseases over the last five years. Determine a baseline number to describe thecurrent extent of the disease in the catchment area.As appropriate, take into account factors for diseases such as malaria or cholerawith seasonal increases.State the threshold clearly as number of cases per month or week, so that healthstaff responsible for surveillance activities can readily recognize when the thresholdis reached.Periodically, revise the epidemic threshold and adjust it accordingly depending onpast and current trends of the disease. If the extent of the disease's burden ischanging (for example, cases are increasing), then adjust the threshold.This indicator can be measured every three monthsby a district supervisor at the health facility level.A proportion of 80% or more on this indicator isregarded as highly satisfactory in terms of thetimeliness of outbreak reporting. When thesurveillance system is good, the indicator shouldapproach 100%.Strengths and LimitationsPrecise measurement of this indicator depends onhealth facilities having the capacity to detectsuspected outbreaks of epidemic-prone diseasesin the district accurately (i.e. the denominator).If the total number of suspected outbreaks ofepidemic-prone diseases in the district isunderestimated, the measurement mayoverestimate performance no matter howpromptly the outbreaks are notified to theprovinces. Depending on the comprehensivenessof the surveillance system, the indicator may ormay not include reporting from private healthfacilities or nongovernmental organizations.176Chapter 5

PROPOROPORTIONOF DISISTRICTTRICTS WITH CURRENTTRENDANALALYSIS(LINEGRAPHSRAPHS) FORSELECTEDPRIORITRIORITYDISEASESCore e IDSR Qualituality y IndicatorDefinitionProportion of districts with current trend analysis(line graphs) for selected priority diseases at a giventime (e.g. month).Numerator: Number of districts with currenttrend analysis (line graphs) for selected prioritydiseases at a given time.Denominator: Total number of districts visitedat same point in time.“Priority diseases:” The list of priority diseasescould vary from country to country, depending onthe local epidemiological situation. WHOrecommends the following nineteen prioritydiseases:(1) Epidemic-prone diseases such as cholera,diarrhea with blood (shigella), measles,meningitis, plague, viral hemorrhagicfevers, and yellow fever;(2) Diseases targeted for eradication andelimination such as acute flaccid paralysis(AFP)/polio, dracunculiasis, leprosy, andneonatal tetanus;(3) Other diseases of public health importancesuch as pneumonia in children less thanfive years of age, new AIDS cases, malaria,onchocerciasis, sexually transmittedinfections (STIs), trypanosomiasis, andtuberculosis (WHO/CDC, 2001).Measurement ToolsSupervisory visit reports (see Annex 5.6 on page204 for a sample supervisory checklist forsurveillance and response activities at the healthfacility level)Integrated Disease Surveillance and ResponseWhat It MeasuresThis indicator measures the extent to whichdistricts are analyzing the surveillance data on aroutine basis for each specific priority disease.Trend analysis of disease cases and deaths over timecan provide key information during an outbreakto enable the identification of the most appropriateand timely control actions to limit the outbreakand prevent further cases from occurring. Analysisof trend data can also facilitate the detection ofabrupt or long-term changes in disease occurrenceand provide information to enable improvementin district public health activities that targetdiseases such as malaria, tuberculosis, HIV/AIDSand vaccine preventable diseases. These diseasescan account for up to 80% of the deaths due topriority diseases and conditions, many of whichoccur in children less than five years of age(WHO-AFRO and CDC, 2001).How to Measure ItThe data for measuring this indicator can beobtained by reviewing an analysis book, if one ismaintained at the district level, during asupervisory visit. The availability of a line graphshould be assessed for each priority diseasespecified by national policy. A district is includedin the numerator if current trend analysis isavailable for each priority disease specified bynational policy. If the quality of the surveillancesystem is satisfactory, this indicator shouldapproach 100%. This indicator can be measuredevery three months by the provincial supervisorthrough his/her routine supervisory visit reports.Strengths and LimitationsThe strength of this indicator is that it can bemeasured for a specific priority disease and at thehealth facility level. It should be noted, however,that while the indicator measures the capacity of177

district health offices (or health facilities) toanalyze data on priority diseases, it does notprovide a measure of the quality of the datacollected and used for line graphs. For example,the indicator does not assess whether healthworkers use standard case definitions to identifyand record suspected cases of priority diseases seenin their health facilities (reliability), or whetherthe condition as reported refer the true conditionas it occurs (validity). While calculation of theindicator is straightforward and not laborintensive, the act of collecting data for thenumerator and denominator is not withoutsignificant costs in terms of travel, materials, time,or installing electronic data management systemsat sub-national levels.Note that health facilities are also expected to dotrend analysis. The indicator can be modified toapply to health facilities and data for measuringthe indicator collected during routine supervisoryvisits:• Proportion of health facilities with currenttrend analysis (line graphs) for selected prioritydiseases at a given time.178Chapter 5

PROPOROPORTIONOF REPOREPORTS OF INVESNVESTIGTIGATEDTED OUTBREAKUTBREAKSTHAT INCLCLUDECASEASE-B-BASEDDATA RECECORDEDANDANALALYZEDCore e IDSR Qualituality y IndicatorDefinitionProportion of reports of investigated outbreaksthat include case-based data recorded andanalyzed.Numerator: Number of outbreak investigationreports that include case-based data recordedand analyzed (includes two-page IDSRoutbreak report, an epi-curve, map, persontables, and line lists or case-based formsattached to the outbreak report) in a given timeperiod (for example, during the last threemonths).Denominator: Total number of outbreakinvestigation reports in the same period.“Outbreak:” An epidemic limited to localizedincrease in the incidence of a disease, e.g., in avillage, town, or closed institution.Measurement ToolsOutbreak investigation reports; case reportingforms; line lists; computerized HIS/MIS databasesWhat It MeasuresA primary feature of a surveillance system is toinvestigate and confirm reported cases andoutbreaks of diseases. Recording and analysis ofcase-based data is important in order to identifythresholds to take action both for routinelyreported priority diseases (diseases of public healthimportance) and case-based diseases (epidemicpronediseases, and diseases targeted foreradication or elimination). Case-based data areessential for planning outbreak response becausethey describe precisely the population at risk fortransmission of the disease or condition, thegeographic extent of the problem, clusters orpatterns of transmission or exposure, and whenexposure to the agent that caused the illnessoccurred. This indicator measures the extent towhich outbreak investigation reports include casebasedinformation.How to Measure ItMeasurement of this indicator requires a reviewof reports of investigated outbreaks to assesswhether they contain the following informationfor suspected and confirmed cases: age group,gender, urban or rural residence, immunizationstatus, inpatient and outpatient status, risk factors,outcome of the episode (for example, whether thepatient survived, died, or his/her survival status isunknown), laboratory results, final classificationof the case, and other variables relevant to thedisease. The outbreak investigation report shouldnot only include this information but also an epicurve,map, person tables, and line lists or casebasedforms in order to be counted in thenumerator. These data are derived from registerreviews, case reporting forms, and line lists thatare used to record information about new cases inhealth facilities or in searches of the community.In some countries, districts have the overallresponsibility for investigating outbreaks. In othercountries, large health facilities will undertakesome or all aspects of investigating outbreaks forsome diseases or conditions. These guidelinesassume that the district is responsible for leadingthe investigation.This indicator should be measured every threemonths by a provincial supervisor at the districtlevel.Integrated Disease Surveillance and Response179

Strengths and LimitationsThis indicator does not provide a measure of thetimeliness and completeness of outbreakinvestigation. CDC/WHO recommends thatdistricts investigate suspected outbreaks within 24hours of notification. Depending on thecomprehensiveness of the surveillance system, theindicator may or may not include reporting fromprivate health facilities or nongovernmentalorganizations. A lack of change in the indicatorwould be difficult to interpret due to its compositenature. In this case, disaggregating the indicatorinto its component parts (a two-page IDSRoutbreak report; an epi-curve map; person tables;attachment of line lists or case based forms to theoutbreak report) would enable the identificationof where problems lie in the recording and analysisof case-based data.180Chapter 5

-PPROPOROPORTIONOF OUTBREAKUTBREAKS OF EPIDEMICPIDEMIC-P-PRONEDISEASESWITH LABABORAORATORORY RESULESULTSCore e IDSR Qualituality y IndicatorDefinitionProportion of outbreaks of epidemic-pronediseases with laboratory results during a specifiedperiod.Numerator: Number of outbreaks of epidemicpronediseases with laboratory results during aspecified period.Denominator: Total number of outbreaks ofepidemic-prone diseases needing laboratoryconfirmation in the same period.“Outbreak:” An epidemic limited to localizedincrease in the incidence of a disease, e.g., in avillage, town, or closed institution.“Epidemic-prone diseases:” These include cholera,diarrhea with blood (shigella), measles, meningitis,plague, viral hemorrhagic fever, and yellow fever.“Specified period:” The time frame may varydepending on the country’s application of theindicator.Measurement ToolsAdministrative records (case-based surveillancereporting forms); supervisory visit reports;laboratory reports/resultsWhat It MeasuresThis indicator measures the extent to whichlaboratory testing is used to support or confirmsuspected outbreaks. Because there are severaldiseases or conditions with similar signs andsymptoms as other diseases or conditions, havinglaboratory support for a diagnosis increases theIntegrated Disease Surveillance and Responselikelihood that the diagnosis is correct and thatpublic health action will be efficient andappropriate (WHO-AFRO and CDC, 2001). Inorder to perform this surveillance function, alaboratory network in districts needs to be in placeand functional. WHO (2001) has developed ageneric laboratory assessment tool that can be usedto rapidly assess the functional laboratory capacityfor diagnosis of priority diseases for surveillanceat all levels of the health system as part of theoverall assessment of national surveillance systems.Components of laboratory functioning arehighlighted in Figure 5.3 on the following page.Proper specimen collection, storage, handling, andtransport are also a critical factor affectinglaboratory performance.How to Measure ItThis indicator is measured by reviewing case-basedsurveillance reporting forms used to recordinformation about individual cases as well aslaboratory results and information about thetimeliness of laboratory testing. A checklist forsupervising surveillance and response activities atthe health facility may be used (see Annex 5.6 onpage 217). A value of 80% or higher on thisindicator is regarded as highly satisfactory in termsof availability of laboratory confirmed results.When the surveillance system is good, theindicator should approach 100%. Countries maychoose to adapt the recommended target and makeit more incremental as performance problems areaddressed.Strengths and LimitationsThis indicator is simple to calculate. The datarequired for the numerator and denominator canbe collected as a part of the routine supervisoryvisits made by provincial officers. Whilecalculation of the indicator is straightforward andnot labor intensive, the act of collecting data for181

Figure 5.3. The role of laboratories in a well-functioning surveillance system.For the laboratory component of a disease surveillance system to function well, the followingare required:X One good national reference laboratory (NRL) performing biological diagnosis, including the following:• Premises are adequate, including adequate electricity, water, and benches.• Staff is well trained and motivated.• Equipment is present, fully functional, and somebody knows how to maintain and fix it.• There is no shortage of small material, consumables, and reagents.• It is fully linked to the international public health laboratory and proficiency testing networks.X One good NRL performing biological diagnosis, including the following:• Premises are adequate, including adequate electricity, water, and benches.• Staff is well trained and motivated.• Equipment is present, fully functional, and somebody knows how to maintain and fix it.• There is no shortage of small material, consumables, and reagents.• It is fully linked to the international public health laboratory and proficiency testing networks.• Provide some tools to epidemiologists; AST, serotype, virulence factors.• Have a good laboratory network that provides samples from remote areas.• Provide, if necessary, a mobile intervention team that can be displaced to a critical area.X Communication roles of the NRL should be the following:• As a disease reporting center, be strongly linked to disease surveillance authorities.• Provide data from remote areas.• Organize and transmit data and comments to disease surveillance authorities.• Drive regular supervision, quality control activities, and training workshops for the laboratorynetwork.Source: Ministry of Labor, Health and Social Affairs and National Center for Disease Control. 2002. Assessment of VaccinePreventable Diseases Surveillance Systems in Georgia. Technical Report No. 028. Bethesda, MD: The Partners for Health ReformplusProject, Abt Associates Inc., p. 25.182Chapter 5

the numerator and denominator may entailsignificant costs in terms of travel, materials, time,or installing electronic data management systemsat sub-national levels.The indicator can be measured at the district,provincial/regional, and national level. Precisemeasurement of this indicator relies on healthfacilities having the capacity to detect outbreaksof epidemic-prone diseases accurately (i.e., thedenominator). If the total number of outbreaksof epidemic-prone diseases in a district isunderestimated, the assessment of the extent oflaboratory confirmation may be overestimated.Depending on the comprehensiveness of thesurveillance system, the indicator may or may notinclude reporting from private health facilities ornongovernmental organizations.It should be noted that the numerator does notmeasure the specificity of the lab results. Manyfactors can affect the reliability of interpretation of alaboratory test result, including inappropriate serumcollection, delay in transportation or refrigeration,and inadequate storage media.While it is important that surveillance systems arelinked to a continuum of laboratories providing themost basic to the most sophisticated services, it isunlikely that resources are sufficient everywhere toprovide the entire continuum.Integrated Disease Surveillance and Response183

PROPOROPORTIONOF CONFIRMEDOUTBREAKUTBREAKS WITHRECECOMMENDEDRESPONSECore e IDSR Qualituality y IndicatorDefinitionProportion of confirmed outbreaks withrecommended response.Numerator: Number of confirmed outbreakswith recommended response during a specifiedperiod.Denominator: Total number of confirmedoutbreaks in same period.“Confirmed outbreak:” An outbreak in which casesthat meet the clinical case definition have beenconfirmed through laboratory diagnostic testing orthrough epidemiological linkage to a laboratoryconfirmed case.“Recommended response:” For epidemic-prone diseaseand for diseases targeted for elimination, a confirmedcase should trigger a response action such asconducting an immunization activity, enhancingaccess to safe drinking water, community campaigns,and improving case management. For other prioritydiseases of public health importance, a confirmedoutbreak should prompt an appropriate response suchas improving coverage for specified immunizations,strengthening case management for IMCI diseases,providing information, education andcommunication about preventing and controlling thedisease, and so on. Table 5.3 on the following pagelists recommended responses to confirmed outbreaksof selected common childhood illnesses and vaccinepreventablediseases. See Section 8 of CDC/WHO(2001) for further details on disease-specificepidemic-response activities.“Specified period:” The time frame may varydepending on the country.Measurement ToolsSupervisory visit reports; outbreak investigation; andresponse reportsWhat It MeasuresWhen an outbreak is confirmed, or a need for publichealth action is identified, an appropriate responseis required based on outbreak investigation resultsand data analysis conclusions. This indicatormeasures whether the health system is functioningeffectively to control outbreaks.How to Measure ItThis indicator is measured by the provincialsupervisor every three months at the district levelthrough his/her routine supervisory visit reports.Each confirmed outbreak is counted in thenumerator if it was responded to according to thedisease-specific guidelines. Disease-specific responsechecklists may be prepared (according to WHO/CDC guidelines) for ease of measurement. A valueof 80% or higher on this indicator is regarded ashighly satisfactory in terms of appropriate responseto confirmed outbreaks. When the surveillancesystem is good, the indicator should approach 100%.Strengths and LimitationsOne advantage of this indicator is that it can bemeasured at the national and provincial/regional level.While the calculation of the indicator isstraightforward and not labor intensive, the act ofcollecting data for the numerator and denominatormay be costly in terms of travel, materials, time, orinstalling electronic data management systems at subnationallevels. It is to be noted that the indicatordoes not measure whether public health actions totreat and control the disease were taken in a timelymanner. The indicator also assumes that necessaryemergency response funds and resources are available.Integrated Disease Surveillance and Response185

Table 5.3. Recommended responses to confirmed outbreaks for selected diseases that are of relevanceto child health programs186Name of Priority DiseaseDiarrhea withdehydration in childrenless than five years of ageMalariaMeaslesMeningitisNeonatal tetanusRecommended Response for Confirmed OutbreakIf the number of cases or deaths increase to two times the number usually seen ina similar period in the past:• Assess health worker practice of IMCI guidelines for managingcases and improve performance for classifying diarrhea withdehydration in children less than five years of age.• Teach mothers about home treatment with oral rehydration.• Conduct community education about boiling and chlorinatingwater, and safe water storage and preparation of foods.If the number of new cases exceeds the upper limit of cases seen in a previous nonepidemicperiod in previous years:• Evaluate and improve, as needed, prevention strategies, such as useof permithirin-impregnated bed nets, especially for young children,pregnant women, and other high-risk populations.If an outbreak is confirmed:• Improve routine vaccine coverage through the EPI, and leadsupplemental vaccination activities in areas of low vaccine coverage.• Mobilize the community early to enable rapid case detection and treatment.Respond to action threshold:• Begin mass vaccination campaign.• Distribute treatment supplies to health centers.• Treat according to epidemic protocol.• Inform the public.• Define the age group at highest risk (usually persons age onethrough 30 years of age) and complete a mass vaccination campaignwithin 10 days of outbreak detection.• Mobilize the community to permit early case detection and treatment,and improve vaccine coverage during mass vaccination campaignsfor outbreak control.If a case is confirmed through investigation:• Immunize the mother with at least two doses of tetanus toxoid andother pregnant women in the same locality as the case.• Conduct a supplemental immunization activity for women ofchildbearing age in the locality.• Improve routine vaccination coverage through EPI and maternalimmunization program activities.• Educate birth attendants and women of childbearing age on the needfor clean cord cutting and care. Increase the number of trained birthattendants.Chapter 5

Table 5.3. Recommended responses to confirmed outbreaks for selected diseases that are of relevanceto child health programs (continued)Name of Priority DiseaseRecommended Response for Confirmed OutbreakPneumoniaPoliomyelitis (acuteflaccid paralysis)Yellow FeverIf the number of cases or deaths increases to two times the number usually seen duringa similar period in the past:• Assess health worker practices of IMCI guidelines for assessing,classifying, and treating children with pneumonia and severepneumonia.• Identify high-risk populations through analysis of person, place, and time.• Conduct community education about when to seek care for pneumonia.If a case is confirmed:• If wild polio virus is isolated from stool specimen, refer to nationalpolio eradication program guidelines for recommended actions. Thenational level will decide which actions to take and may include:® Specify reasons for non-vaccination of each unvaccinated caseand address the identified deficiencies.® Immediately conduct “mopping-up” vaccination campaignaround the vicinity of the case.® Conduct surveys to identify areas of low OPV coverage duringroutine EPI activities, and improve routine vaccine coverage ofOPV and other EPI antigens.® Lead supplemental vaccination campaigns during NationalImmunization Days (NIDs) or Sub-National ImmunizationDays (SNIDs). Focus supplemental vaccination activities in areasof low vaccine coverage during EPI. Consider use of house-tohousevaccination teams in selected areas.If a single case is confirmed:• Mobilize the community early to enable rapid case detection and treatment.• Conduct a mass campaign in the appropriate age group in the area (agessix months and older) and in areas with low vaccine coverage.• Identify high-risk population groups and take steps to reduce exposureto mosquitoes.• Improve routine and mass vaccination campaigns to include yellow feverin high-risk areas.Source: WHO-AFRO and CDC, 2001.Integrated Disease Surveillance and Response187

ATTTTACK RATEPriorioritity y Indicator for SururveileillancelanceDefinitionThe number of new cases of a specified diseaseper 100,000 population at risk of the disease in agiven period.Numerator: Number of new cases of aspecified disease during a given period x100,000.Denominator: Total population at risk of thedisease at the start of the same period.“Case:” A person who has the particular disease,health disorder, or condition which meets the casedefinitions for surveillance and outbreakinvestigation purposes.Measurement ToolsData from field investigationWhat It MeasuresThe attack rate is a measure of disease frequencyapplied to a narrowly defined population observedfor a limited period of time, such as during anepidemic. An attack rate is used when theoccurrence of disease among a population at riskincreases dramatically over a short period of time.How to Measure ItThis indicator is a simple ratio that isstraightforward and easy to calculate. It can bemade age-specific by using the following foursteps:(1) Calculate the number of persons who are inthe age group in the area (e.g., persons aged0-4 years);(2) Divide 100,000 by the number of persons inthe age group;(3) Tally the number of cases in the age group forthe period of time chosen; and4) Multiply the result of Step 2 by the numberof cases in the age group.The result is the age-specific attack rate. Thedifficulty of calculating the attack rate based ondata from field investigations stems from problemsrelated to estimating the size of the totalpopulation at risk in a given age group. These are,in turn, the result of inaccurate or outdated censuscounts, population migrations, and unforeseenchanges in birth rates or infant mortality. Wheredata on the number of persons in a given age groupare unavailable, evaluators can estimate thesenumbers from the total population of the area ordistrict and the typical age distribution for therelevant region of the world. The box on the nextpage illustrates how to estimate the totalpopulation for a given age group using thisprocedure.Integrated Disease Surveillance and Response189

For example:Strengths and LimitationsThe attack rate is useful in comparing the risk ofdisease in groups with different exposures. It canbe used to plan disease prevention activities (e.g.,a vaccination strategy) and to target preventionactivities to the groups with the highest attackrates. As the period of time over which themeasurement is based is completely arbitrary, theattack rate is a versatile indicator.190Despite the versatility of an arbitrary time period,the value of the indicator is only as good as thedata that are used for its calculation. Thus, datawhich are incomplete, confounded, or inaccurate,will lead to attack rates that may not berepresentative of the effectiveness of diseaseprevention activities.Chapter 5

CASEFATALITALITY RATEFOROUTBREAKUTBREAKS OF EPIDEMICPIDEMIC-PRONEDISEASESCore e IDSR Qualituality y Indicator(with adaptations for child health)DefinitionProportion of persons diagnosed as having aspecified epidemic-prone disease who die fromthat disease within a given period. The case fatalityrate is usually expressed as a percentage.Numerator: The number of deaths of personsdiagnosed as having a specified epidemicpronedisease in a given period.Denominator: Total number of personsdiagnosed as having the specified epidemicpronedisease in the same period.“Epidemic-prone diseases:” These include cholera,diarrhea with blood, measles, meningitis, plague,viral hemorrhagic fever, and yellow fever.The definition and calculation of the indicator canbe adapted to refer to children aged 0-59 months,as follows:Proportion of children aged 0-59 monthsdiagnosed as having a specified epidemic-pronedisease who die from that disease within a givenperiod.Numerator: The number of deaths of childrenaged 0-59 months diagnosed as having aspecified epidemic-prone disease in a givenperiod.Denominator: The number of children aged0-59 months diagnosed as having the specifiedepidemic-prone disease in the same period.Measurement ToolsClinic registers; patient charts; surveillancereporting forms (see Annex 5.7 for an example ofa monthly surveillance reporting form foroutpatient and inpatient cases and deaths)What it MeasuresThe case fatality rate is a measure of severity ofillness. The indicator aims at measuring progresstowards the reduction of mortality from epidemicpronediseases at the health facility level. Itexpresses the likelihood that a person with aspecific disease will live after entering the healthfacility. A case fatality rate helps to indicatewhether a case is identified promptly, and anyproblems with case management once the diseasehas been diagnosed. It also helps to identify amore virulent, new, or drug-resistant pathogen andindicate poor quality of care or no medical care.Some disease control recommendations forspecific diseases include reducing the case fatalityrate as a target for measuring whether the outbreakresponse has been effective.How to Measure ItThe data for calculating this indicator can bederived from a review of clinic registers or patientcharts or through periodic reporting of data onsuspected cases of the specific disease (cases thatmeet the clinical case definition) from healthfacilities to districts, or from districts to theprovincial/regional level. The indicator divides thetotal number of all deaths from a specific disease(for example, ARI, diptheria, pertussis, measles,neonatal tetanus, etc.) during a specific period bythe total number of reported cases of the diseasein the same period, and multiplying the answerby 100.Integrated Disease Surveillance and Response191

Each epidemic-prone disease has a specific casefatalityrate (CFR) threshold. For example, in thecase of cholera, a CFR greater than 1% indicates,among other things, poor case management. Inthe case of meningitis, a CFR less than or equalto 10% indicates an acceptable level (i.e., the targetlevel), a CFR greater than 25% is considered high,and a CFR lower than 5% is considered very low(i.e., it may be that illness that is not due tomeningitis is being classified as meningitis – “overdiagnosis”– or that severely ill patients are notreaching health facilities). In the case of measles,a CFR above 4% is considered high mortality(WHO, 2003b).Strengths and LimitationsOnce data on cases are being collected, thisindicator is relatively easy to calculate. It can alsorespond to changes over a relatively short period,for example, 6-12 months. This indicator mostlyhelps service management at the level of eachfacility. This indicator helps to: (1) indicatewhether a case is identified promptly; (2) indicateproblems with case management once the diseasehas been diagnosed; (3) identify a more virulent,new, or drug-resistant pathogen; (4) indicate poorquality of care or no medical care; and (5) comparethe quality of case management between differentcatchment areas, health facilities, or districts.One way to disentangle the components of theCFR is to gather information on other indicatorsof the quality of care, such as the admission-totreatmenttime interval. Another approach wouldbe to gather information about the condition ofthe child at the time of admission. This couldhelp disentangle the effect of patients’ conditionfrom that of quality of care.It may not be valid to compare case fatality ratesbetween facilities, especially between healthcenters and hospitals, since children with seriousillness could be referred to the hospital at the lastmoment, where they may die. This would lowerthe CFR at the health center and raise it at thehospital. Thus, interpretation requires comparingthe CFR for a particular facility over time and notcomparison between facilities. When using thisindicator to monitor trends over time in the qualityof care, one caveat worth mentioning is that datafrom a recent year may be incomplete if there is asignificant lag time in reporting.When interpreting this indicator, one shouldconsider that it is sometimes difficult todistinguish deaths from a particular disease fromdeaths from other causes. Thus, the numeratorcan be as affected by errors in diagnosis, as bychanges in classification. The case fatality rate isalso affected by the quality and promptness ofmedical care provided in the facility, the conditionof the child upon arrival, and distance from thehealth facility. Case-fatality rates may beunderestimates because of incomplete reportingof deaths. For example, a CFR under 5% foryellow fever may suggest an epidemic is justbeginning, or “over-diagnosis,” or the fact thatseverely ill cases may not be reaching healthfacilities.192Chapter 5

RefererencesencesCDC. (1997). Manual for the Surveillance of Vaccine-Preventable Diseases. Atlanta: US Department ofHealth and Human Services. CDC:15.1-10.CDC. (2001). Updated Guidelines for Evaluating Public Health Surveillance Systems: Recommendationsfrom the Guidelines Working Group. MMWR, 50(No. RR-13):11-24.CDC. (2003). Integrated Disease Surveillance and Response, Vol. 1, Summer 2003.Gordis, L. (2000). Epidemiology, 2nd Ed. W.B. Saunders Company: Philadelphia.Ministry of Labor, Health and Social Affairs, and National Center for Disease Control. (2002).Assessment of Vaccine Preventable Diseases Surveillance Systems in Georgia. Technical Report No. 028.Bethesda, MD: The Partners for Health Reformplus Project, Abt Associates, Inc.Romaguera RA, German RR, and Klaucke DN. (2000). Evaluating Public Health Surveillance. In:Teutsch SM, and Churchill RE, eds. Principles and Practice of Public Health Surveillance, 2nd Ed. NewYork: Oxford University Press; 176-193.Wharton M, and Ching PL. (2002). Surveillance Indicators. In: VPD Surveillance Manual, 3rd ed;15-1–15-10.White ME, and McDonnel SM. Public Health Surveillance in Low- and Middle-Income Countries.In: Teutsch SM, and Churchill RE, eds. Principles and Practice of Public Health Surveillance, 2nd Ed.New York: Oxford University Press; 287-315.WHO. (1998). WHO-Recommended Standards for Surveillance of Selected Vaccine-Preventable Diseases.Geneva, Switzerland: World Health Organization. WHO/EPI/GEN/98.01.WHO. (2000). Maternal and Neonatal Tetanus Elimination by 2005: Strategies for Achieving andMaintaining Elimination. Geneva, Switzerland: World Health Organization. WHO/V&B/02.09.WHO. (2001). Protocol for the Assessment of National Communicable Disease Surveillance and ResponseSystems: Guidelines for Assessment Teams. Geneva, Switzerland: World Health Organization. WHO/CDS/CSR/ISR/2001.2.WHO. (2002). Accelerating Immunization in the African Region. Retrieved from http://www.afro.who.int/press/2002/pr20021008-12rc03.htmlWHO. (2003a). Communicable Diseases Epidemiological Report. Retrieved from http://www.afro.who.int/csr/ids/bullettins/afro/jan2003.pdfIntegrated Disease Surveillance and Response193

WHO. (2003b). Current Epidemiological Situation of Reported Epidemic Prone Diseases in the Region(Year 2003). Retrieved from http://www.afro.who.int/csr/epr/epidemic_prone_diseases/weekly_feedback1-7_2003.pdfWHO-AFRO/CDC. (2001). Technical Guidance for Integrated Disease Surveillance and Response in theAfrican Region. Retrieved June 20, 2003, from http://www.cdc.gov/epo/dih/Eng_IDSR_Manual_01.pdf194Chapter 5

Annex 5.1. WHO-recommended case definitions for reporting selected suspected priority diseasesfrom the health facility to the districtIntegrated Disease Surveillance and Response195

Annex 5.1. WHO-recommended case definitions for reporting selected suspected priority diseasesfrom the health facility to the district (continued)Source: WHO/CDC (2001), Annex 2, 29-31.196Chapter 5

Annex 5.2. WHO-recommended types of surveillance for vaccine-preventable diseasesIntegrated Disease Surveillance and Response197

Annex 5.2. WHO-recommended types of surveillance for vaccine-preventable diseases (continued)Source: WHO. 1998. WHO-Recommended Standards for Surveillance of Selected Vaccine-Preventable Diseases. WHO/EPI/GEN/98.01. Geneva: World Health Organization.198Chapter 5

Annex 5.3. Sample form for recording timeliness and completeness of monthly reporting from thehealth facility to the districtSource: WHO and CDC (2001), Annex 29, p. 185.Integrated Disease Surveillance and Response199

Annex 5.4. Generic case-based reporting form – from health facility/worker to district health team200Chapter 5

Annex 5.4. Generic case-based reporting form – from health facility/worker to district health team(continued)If Lab Specimen CollectedSource: WHO and CDC (2001), Annex 8, pp. 57-58.Integrated Disease Surveillance and Response201

Annex 5.5. Generic line list – for reporting from health facility to district and for use during outbreaks202Chapter 5

Annex 5.5. Generic line list – for reporting from health facility to district and for use during outbreaks(continued)Source: WHO and CDC (2001), Annex 9, pp. 62-63.Integrated Disease Surveillance and Response203

Annex 5.6. Sample supervisory checklist for surveillance and response activities at the health facility204Chapter 5

Annex 5.6. Sample supervisory checklist for surveillance and response activities at the health facility(continued)Source: WHO and CDC (2001), Annex 31, pp. 188-189.Annex 5.7. Monthly surveillance summary report form for out-patient cases and in-patient cases anddeaths (district to next level)Chapter 5. Integrated Disease Surveillance and Response205

6INTEGRATEDMANAGEMENT OFCHILDHOOD ILLNESS(IMCI): IMPROVEDHEALTH WORKERSKILLSIndicators:• Child checked for three danger signs• Child checked for the presence of cough, diarrhea, and fever• Child’s weight checked against a growth chart• Child’s vaccination status checked• Index of integrated assessment of sick child• Child under two years of age assessed for feeding practices• Child needing an oral antibiotic and/or antimalarial is prescribed thedrug(s) correctly• Sick child not needing antibiotic leaves the facility without antibiotic• Caretaker of sick child is advised to give extra fluids and continue feeding• Child needing vaccinations leaves facility with all needed vaccinations• Caretaker of child who is prescribed ORS and/or oral antibiotic and/orantimalarial knows how to give the treatment• Sick child needing referral is referred

HAPTER 6. ICHAPTER6. INTEGRANTEGRATEDTED MANANAGEMENTOF CHILDHOODILLNESSMPROVEDHEALEALTHWORKERSKILLKILLS(IMCI): IMPRIMCI is an integrated programmatic approachfor reducing infant and childhood morbidityand mortality and improving nutritional status,through the provision of timely appropriate carefor illnesses most prevalent among children underfive. Instead of having vertical approaches forevery major childhood disease, as was the case withprevious efforts, IMCI focuses on dealing withcommon childhood illnesses in a more efficientand cost effective manner. IMCI has three majorcomponents, each of which is adapted in countrieson the basis of local epidemiology, health systemcharacteristics, and culture. These componentsinclude: (a) improving the skills of health workersthrough training and reinforcement of correctperformance; (b) improving health system supportfor the delivery of child health services, includingthe availability of drugs, effective supervision andthe use of monitoring and health informationsystem data; and (c) improving family practicesthat are important for children’s health anddevelopment. The latter component alsoencourages the development and implementationfor community and household-based interventionsto increase the proportion of children who areexposed to these practices.Figure 6.1 shows a simplified model that outlinesthe pathways through which the different IMCIinterventions are expected to lead to changes inchild mortality, health, and nutrition (Bryce et al.,2004). The model does not directly addresscontextual factors that could affect the outcomesand impact of IMCI interventions. However, therole of these factors in modifying the effects ofIMCI activities on child mortality, health, andnutrition outcomes or producing effects that areindependent of or simultaneous with IMCIinterventions is recognized.Figure 6.1. Outline of the integrated management of childhood illness (IMCI) impact model.Source: Bryce et al. (2004).Chapter 6. Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills207

This chapter focuses on indicators for monitoringand evaluating health facility-based interventionsintended to improve health worker skills. Thesetraining interventions are based on a set ofguidelines for assessing signs and symptoms;classifying illness and identifying neededtreatment; providing appropriate treatment,referral, and counseling of the child’s caretaker,depending on the classification identified; andproviding follow-up care. The guidelines alsocover the assessment of nutrition andimmunization status and potential feedingproblems, and effective communication with thechild’s caretaker.Table 6.1 presents illustrative processes andoutputs related to interventions to improve healthworker skills. We do not address efforts to improvethe skills of community health workers but focuson the training of first-level public health providersin IMCI case management. IMCI traininginterventions are expected to lead to improvedhealth worker skills and competence as well asreduced missed opportunities for vaccination andtreatment of childhood illness, which should leadto increased patient satisfaction and improvedhome care. Effective counseling of caretakers byIMCI-trained health workers is expected to leadto improvements in early case management,Table 6.1. Illustrative processes, outputs, and outcomes associated with improving health worker skillsProcessesOutputsOutcomesImpact• Trainingfirst-levelhealthworkers anddistrict-levelsupervisorsFunctional Outputs• Health workerstrained in IMCI• District supervisorstrained in IMCIService Outputs• Improved healthworker skills• Improved healthworker competence• Reduced missedopportunities forvaccination andtreatment ofchildhood illness• Increased patientsatisfaction• Improved caretakerknowledge andpracticeso Early casemanagemento Appropriatecare seekingo Compliancewith treatment• Reducedmortality• Improvedchild healthand nutrition208Chapter 6

appropriate home treatment, and compliance withtreatment recommendations. These outcomesshould lead, in turn, to reductions in the casefatality rates of diarrhea, pneumonia, measles,malaria, and other severe bacterial infections, aswell as to reductions in malnutrition.Service Provision AssessmentThe Service Provision Assessment (SPA) is themost comprehensive tool for measuring the qualityof care at the health facility level. The SPA is anational survey of a representative sample of publicand/or private facilities that provide maternal,child, and reproductive health services. The SPAmeasures not only the quality of care but also theoverall functioning of a health facility, as reflectedin the set of questions it addresses:(1) To what extent are the surveyed facilitiesprepared to provide the priority services?(2) To what extent does the service deliveryprocess follow generally accepted standards?(3) To what extent do support systems formaintaining or improving the existing servicesexist, and how well are they functioning?(4) What are the issues the clients and serviceproviders consider relevant to their satisfactionwith the service delivery environment?The SPA uses four different data collectionmethods. The first is an inventory of resourcesand support services, which provides informationon the preparedness of a facility to provide priorityservices at an acceptable standard of quality. Aspart of the inventory (also known as a facilityaudit), interviewers ask staff about theirqualifications, training, perceptions of the servicedelivery environment, and related issues. Thesecond is a provider interview, during whichinterviewers ask health service providers forinformation on their qualifications (training,experience, and continuing education), supervisionthey have received, and perceptions of the servicedelivery environment. The third is observation ofservices provided. The observation assesses theextent to which service providers adhere to servicedelivery standards. The fourth is exit interviewswith clients who have received services. The exitinterview assesses the client's understanding andperceptions of the consultation/examination, aswell as recall of instructions regarding treatmentor preventive behaviors. Recall of key messagesincreases the likelihood that the client (caretaker)will successfully follow treatment or will performthe preventive behaviors that optimize healthyoutcomes.The SPA provides the following information onchild health:• Preparedness to provide good quality services(specifically, immunization services and basicdiagnosis and outpatient treatment of sickchildren)• Adherence to standards for provision ofservices• Client understanding of consultationThe following components of the SPA are of directrelevance to an evaluation of health worker skillsand competence:Staff: What is the qualification of staff whoprovide the service? Have the health workersreceived periodic continuing education onrelevant topics, and how recently has trainingoccurred? Have the health workers received aminimal level of supervision?Process: Do protocols and standards ofpractice for each service meet generallyaccepted quality standards for basic as well asadvance level services at referral facilities? Doproviders adhere to the standards of practicefor service delivery? The process assessedincludes procedures followed, components ofphysical examinations, as well as theinformation exchanged between the healthworker and client (caretaker in the case of childhealth). The SPA assesses if the process duringservice delivery meets the standards; it does notassess if providers correctly diagnose the problem.Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills209

210Client understanding: What informationregarding the consultation, instructions, orfollow-up can the client recall?Other components related to the ability of healthfacilities to deliver quality services are:Facility resources, equipment, and supplies:What specific equipment and supplies areavailable at the various levels of servicedelivery? Do the elements that are requiredto provide the services meet minimumstandards? Are these elements present,functioning, and in the appropriate locationfor use during service provision? Are theresystems for maintaining adequate availabilityof supplies (inventories; appropriate storage;equipment maintenance and repair/replacement systems), and is there evidenceof effectiveness?Systems for evaluating and monitoringservices: Are routine information systems upto-dateand able to provide basic client andservice provision data? Are there systems formonitoring community coverage if that isexpected of the facility?Facility management: Does the facility havebasic management systems in place, and dothey include community representation? Doesthe facility participate in any financingmechanism that affects the cost to thecommunity or client?Service provision environment: Does thefacility collect very basic information aboutproblems the staff thinks should be addressedto improve their working situation andservices? Does the facility collect datarevealing the opinion of clients regardingissues related to satisfaction with theirconsultation and the service deliveryenvironment? The appendix presents asummary of the information available from theSPA on child health.The SPA identifies the strengths and weaknessesof a set of clinical facilities at a given point in time,and if repeated, the data can demonstrate changesover time. If a program is not achieving its desiredoutcome, the SPA may reveal service-relatedreasons for this shortcoming. Data from the SPAdata can be linked to household-level data fromthe DHS to demonstrate that changes(improvements) in the service deliveryenvironment improve outcomes at the populationlevel.Challenges of Monitoring and EvaluatingHealth Worker SkillsObservation and courtesy biasMost of the indicators for monitoring andevaluating health worker skills are measuredthrough observations of sick child consultationsor exit interviews. In observations of sick childconsultation, it is often not possible to keep theprovider ignorant of actual observation. Therefore,it is likely that a health worker may abide by theguidelines more strictly when he or she is consciousof being monitored. Such observation bias maybe reduced if observation teams spend a longerperiod in the health facility, by the end of whichprovider behaviors may become more normative.An indicator may also be affected by "courtesybias" if the information required to measure it iscollected through an exit interview. The caretakermay respond positively about procedures that thehealth worker was supposed to carry out, eventhough they may not have been performed.Courtesy bias is likely to occur if the caretaker hasdoubts about the confidentiality of the exitinterview or if the caretaker has developed apositive interpersonal relationship with the healthworker. The bias can be reduced if interviewersstress that the responses will remain confidentialand that no identifying information will be puton the questionnaires.Complexity of Indicator MeasurementSome of the indicators recommended formonitoring and evaluating health worker skills areChapter 6

composite indicators that combine severalelements. Thus, the indicators are not alwaysstraightforward to measure. For example, correcttreatment of a particular condition requires correctquestions to be asked, correct assessments to beconducted, and correct treatment to be provided.All these steps need to be carefully monitoredthrough observation and compiled into anaggregate index. Explaining a lack of change overtime in a composite indicator poses a majorchallenge as improvements in some componentsmay be masked by deterioration in others.Prioritizing IndicatorsBecause the quality of care is a complex, multifacetedissue, there could literally be hundreds ofindicators to measure it. As collecting data on allthese indicators would be time-consuming, costly,and overwhelming, the question of prioritizingindicators to measure the quality of care is animportant one. From the point of view ofpracticality, WHO (2001) has developed a set ofcore and supplemental indicators for IMCI atfirst-level facilities. If time and budget permit,evaluators may want to augment these indicatorswith the supplemental indicators listed in Annex6.1 on page 239.Maximizing the Use of Available DataThe challenge remains as to how best to use theavailable data to inform decisions about theprovision of high quality child health services orimprovements in the quality of care.Overburdening the system with a great deal ofinformation may prove to be counter-productivein the absence of clear guidance on how programmanagers can use the data for decision-making.Selection of IndicatorsThis chapter presents WHO priority indicatorsfor the assessment of health worker skills at thehealth facility level. These indicators were selectedby the Interagency Working Group on IMCIMonitoring and Evaluation on the basis of theirvalidity, reliability, and feasibility. The indicatorsare the following:Assessment• Child checked for three danger signs• Child checked for the presence of cough,diarrhea, and fever• Child's weight checked against a growth chart• Child's vaccination status checked• Index of integrated assessment of sick child• Child under two years of age assessed forfeeding practicesCorrect Treatment and Counseling• Child needing an oral antibiotic and/orantimalarial is prescribed the drug(s) correctly• Sick child not needing antibiotic leaves thefacility without antibiotic• Caretaker of sick child is advised to give extrafluids and continue feeding• Child needing vaccinations leaves facility withall needed vaccinations• Caretaker of child who is prescribed ORS and/or oral antibiotic and/or antimalarial knowshow to give the treatmentCorrect Management of Severely Ill Children• Sick child needing referral is referredAlthough the scope of this chapter is limited toonly one of the components of the IMCI strategy(i.e., improving health worker skills), for acomplete evaluation of the outcomes and impactof IMCI, all three components must be monitoredand evaluated. If health workers are competentin case management but do not have theequipment, supplies, or drugs needed to managecases correctly, desired child health outcomescannot be achieved. This implies that activitiesaimed at improving the ability of the health systemIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills211

to deliver IMCI need to be monitored andevaluated alongside activities to improve healthworker skills. For example, increased availabilityof equipment, drugs, and supplies and improvedservice organization at health facilities are expectedto lead to reduced waiting times for clients and amore complete assessment of child health careneeds. Readers are referred to Annex 6.1 for a listof indicators for monitoring and evaluatingimproved health system support for the deliveryof child health services.212Chapter 6

CHILDCHECHECKEDFORTHREEDANANGERSIGNSPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children checked for the threegeneral danger signs.Numerator: Number of sick children aged 2-59 months seen who are checked for threedanger signs (is the child able to drink orbreastfeed; does the child vomit everything;has the child had convulsions).Denominator: Number of sick children aged2-59 months seen.Measurement ToolsService Provision Assessment (SPA); HealthFacility Assessment (HFA); supervision checklistWhat It MeasuresThis indicator measures the performance of oneof the tasks associated with the routine assessmentof sick children aged two months to five years:checking for general danger signs. The indicatorpresupposes that the health worker has beentrained in IMCI. Given this assumption, theindicator measures both the adequacy of IMCItraining to impart these skills and the ability ofthe trainees to assimilate and retain theinformation and skills over time.child has had convulsions during the presentillness. It is recommended that health workersnot rely completely on the caretaker’s report ofwhether the child is able to drink or breastfeedbut observe the mother while she tries tobreastfeed or to give the child something to drink.Strengths and LimitationsMeasuring this indicator is straightforward andcan be done in a routine basis during supervisoryvisits. Immediate feedback can be given to thehealth worker to improve future practice. Theindicator can help identify if health workers of aparticular health facility need refresher training.If no improvement in this indicator is seen overtime, program managers may wish to monitorhealth workers’ assessment of the presence of eachdanger sign separately to identify whether aparticular danger sign is not routinely checked.Sample QuestionsSample instructions from a checklist completedby an observer during a SPA are the following:o Record whether a provider asked about orwhether the caretaker mentioned any of thefollowing:(1) Whether the child is unable to drink orbreastfeed at all;(2) Whether the child vomits everything; and(3) Whether the child has had convulsionswith this sicknessHow to Measure ItMeasurement of this indicator requires directobservation of sick child consultations todetermine whether the health worker asked aboutor the caretaker reported each of the followingdanger signs: the child is unable to drink orbreastfeed, the child vomits everything, and theIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills213

CHILDCHECHECKEDFORTHE PRESENRESENCEOF COUGHOUGH,DIARRHEAIARRHEA, AND FEVERPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children aged 2-59 monthschecked for the presence of cough, diarrhea, andfever.Numerator: Number of sick children aged 2-59 months seen whose caretakers were askedabout the presence of cough, diarrhea, andfever.Denominator: Number of sick children aged2-59 months seen.Measurement ToolsSPA; HFA; supervision checklistWhat It MeasuresThis indicator measures health worker compliancewith IMCI clinical guidelines requiring theroutine check of all sick children for the three mainsymptoms that often result in death: cough ordifficult breathing, diarrhea, and fever. Thisenables the health worker to exclude or identifycases of ARI or diarrhea or measles or malaria.The indicator presupposes that health workershave been trained in IMCI. If this assumption ismet, the indicator not only reflects the retentionof skills acquired during IMCI training, but alsomeasures the adequacy of the training to impartthese skills and the ability of the trainees toassimilate and retain the information and skillsover time.How to Measure ItThe data sources for this indicator are checklistsand notes from an expert observation of sick childconsultation. The data required for calculatingthis indicator can also be collected during routinesupervisory visits. A child is counted in thenumerator if the health worker asked the caretakerabout each of the three symptoms. All sickchildren seen at the health facility constitute thedenominator.Strengths and LimitationsThe data required to calculate the indicator aresimple to collect and can easily be incorporatedinto checklists used for the routine supervision ofhealth workers. This indicator can be collectedafter a specific interval post-training (e.g., sixmonths or 12 months) among health workers whoattended an IMCI training course.Sample QuestionsSample questions from the March 2004 versionof the SPA are as follows:o Assessment task: Record whether a providerasked about or whether the caretakermentioned that the child had any of thefollowing major symptoms:(a) Has cough or difficult breathing;(b) Diarrhea; and(c) Fever or body hotness.Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills215

CHILDHILD’S WEIGHTCHECHECKEDAGAINSAINST A GROWTHCHARHARTPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children aged 2-59 months whoare weighed the day they are seen and whose weightsare checked against a recommended growth chart.Numerator: Number of sick children aged 2-59months who are weighed the day they are seenand whose weights are checked against arecommended growth chart.Denominator: Total number of sick children aged2-59 months seen.Measurement ToolsSPA; HFA; supervision checklistWhat It MeasuresThis indicator measures health worker compliancewith IMCI guidelines for the routine assessmentof the nutritional status in sick children. Thereare two main reasons for this assessment. Thefirst is to identify children with severe malnutritionwho are at increased risk of mortality and needurgent referral. Second, the assessment of thenutritional status of sick children helps to identifychildren with growth faltering who may benefitfrom nutritional counseling. All sick childrenshould be assessed for malnutrition.How to Measure ItThe data are gathered by direct observation of sickchild consultations. The observer records whetherthe health worker weighed the child and plotted thechild’s weight on a recommended growth chart(usually a standard WHO or national growth chart).Strengths and LimitationsData for this indicator are easy to collect during sickchild observation or routine supervisory visits. Theygive a good indication of health worker compliancewith IMCI guidelines regarding the nutritionalassessment of all sick children. The indicator can beapplied at a specific interval post-training to thosewho attended IMCI training to evaluate theretention of this particular component of clinicalassessment skills. This may help identify healthworkers who need refresher training or health centersin which weighing of sick children and recordingthe weight on a growth chart are not enforced.Limitations to the use of observation for measuringquality of sick child assessment have already beendiscussed. These include “observation bias,” in thata health worker may abide to the guidelines morestrictly when he or she is conscious of beingmonitored. Another limitation of the indicatorpertains to variability between observers inmeasurement. This is hard to measure but can beassessed by having two independent observers rate asick child consultation and then comparing thedegree of agreement or disagreement in their ratings.Note that the indicator does NOT reflect the correctmeasurement of the child’s weight, whether thechild’s weight was accurately plotted on a growthchart, how effectively health workers interpreted theinformation on the growth chart, or whether thehealth worker took an appropriate course of actionbased on insights from the growth chart. It is alsodifficult to tell from the indicator whether healthworkers are weighing/not weighing the children atall, or whether children are weighed but their weightsare not plotted on a growth chart.Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills217

CHILDHILD’S VACCINCCINATIONSTATUSTUS CHECHECKEDPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children who have theirvaccination status checked.Numerator: Number of sick children seen whohave their vaccination card or history checked.Denominator: Number of sick children seen.Measurement ToolsSPA; supervision checklistWhat It MeasuresThis indicator is a measure of missed opportunitiesfor vaccinating sick children who are brought tothe health facility.How to Measure ItThe data requirements for calculating thisindicator are the number of sick children who havetheir vaccination card or history checked and thenumber of sick children seen. The data arecollected by observation of sick child consultations.The indicator can also be collected throughchecklists used for routine supervision. Theobserver/supervisor records on a checklist whetherthe health worker checked the child’s vaccinationstatus by reviewing the child’s vaccination card orasking the caretaker questions about the child’svaccination history.Strengths and LimitationsThe strength of this indicator lies in its scope forassessing missed opportunities for vaccinating sickchildren who present to the health facility. Theimmunization status of every sick child broughtto a health facility should be checked. Theindicator only measures whether health workersare screening the vaccination status of sick infantsand children, not whether they act on thisinformation appropriately by vaccinating childrenin need of a vaccination on the same day orreferring them to the next vaccination session.Sample QuestionsThe MEASURE DHS+ SPA (version datedMarch 2004) observation checklist for sick childconsultation at health facilities includes thefollowing instructions:o Record whether a provider asked about orperformed other assessments of the child’shealth by doing the following:• Look at the child’s immunization card orask the caretaker about the child’svaccination history?Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills219

INDEXOF INTEGRANTEGRATEDTED ASSESSMENTOF SICICKCHILDRENPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionArithmetic mean of assessment tasks performedper sick child assessed.The index is made up of the following 10assessment tasks that must be taken by the healthprovider to check for general danger signs, mainsymptoms, nutritional status, and immunizationstatus:• Sick child checked for ability to drink orbreastfeed (1 point)• Sick child checked on whether he/she vomitseverything (1 point)• Sick child checked for presence of convulsionsduring present illness (1 point)• Sick child checked for presence of cough andfast/difficult breathing (1 point)• Sick child checked for presence of diarrhea(1 point)• Sick child checked for presence of fever(1 point)• Sick child weighed on the day of the visit tothe health facility (1 point)• Sick child’s weight plotted against arecommended growth chart on the day of thevisit to the health facility (1 point)• Sick child checked for palmar pallor (1 point)• Sick child’s vaccination status checked(card or history) (1 point)Measurement ToolsSPA; HFAWhat It MeasuresCorrect assessment is required in order to classifyand treat children appropriately. This indicator isan aggregate measure of the quality of assessmentof sick children at the facility level. It reflects theextent to which health providers perform the stepsrequired to correctly assess sick children. Thesesteps include: (1) checking for general dangersigns; (2) checking main symptoms; (3) checkingnutritional status; and (4) checking immunizationstatus.How to Measure ItThe data for calculating this indicator are derivedfrom a questionnaire completed by an expertobserver to evaluate health worker performanceduring sick child consultations. Each routine taskmentioned in the definition is assigned “1” point.To calculate this indicator, the following data areneeded for each health provider that manages sickchildren:• The number of assessment tasks completedfor each sick child seen by a health provider• The total number of sick children seen by thathealth provider• The sum of assessment tasks completed by thathealth provider for all sick children seenThe numerator is the sum of assessment taskscompleted by a health provider for all sick childrenseen. The denominator is the total number of sickchildren seen by the health provider. The indexranges from zero (indicating no routine assessmenttasks were completed for any sick child seen) to10, signifying that all routine assessment tasks werecompleted for all sick children seen by a healthworker. The index may be converted into a grade.Grades could range from weak through moderateand strong, depending on the index’s range.Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills221

Strengths and LimitationsThis summary indicator may be useful fordemonstrating change in health workers’performance of assessment tasks over time, sincehealth workers may progressively increase thenumber of essential tasks that are completed. Theindex can be aggregated across health workers ina facility to produce an average sick childassessment score for health facilities. The indexcan also be aggregated across facilities in a districtto produce an index of integrated assessment atthe district-level. If the index is used to comparethe quality of sick child assessment across healthfacilities or districts, the following two indicatorscould be used: (1) the average number ofassessment tasks per sick child and (2) thepercentage of health workers performing aminimum number of assessment tasks. Thecomparison of the performance of health workersamong different health facilities in a particulardistrict, or among different districts, can helpidentify health facilities or districts that areperforming below standards.Questions have been raised about the utility of acomposite score, as it could mask deterioration inthe performance of some assessment tasks andimprovements in others. For program planningand monitoring, it may be useful to simply trackeach assessment task comprising the indexseparately. This would enable the identificationof which tasks are not usually performed.Individual screening tasks can then be targetedthrough routine supervision or refresher training.222Chapter 6

CHILDUNDERTWO YEARSOF AGEASSESSEDFORFEEDINEEDING PRARACTICESPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children under two years of agewhose caretakers are asked about breastfeeding,complementary foods, and feeding practicesduring this episode of illness.Numerator: Number of sick children undertwo years of age whose caretakers are asked ifthey breastfeed this child, whether the childtakes any other food or fluids other than breastmilk, and if during this illness the child’sfeeding has changed.Denominator: Number of sick children undertwo years of age seen.Measurement ToolsSPA; HFA; supervision checklistWhat It MeasuresThis indicator is a composite of the steps a healthworker must undertake in order to correctly assessa sick child’s feeding and counsel the mother tosolve any feeding problems that exist. Thus, itmeasures the correct assessment of a sick child’sfeeding. The IMCI guidelines require that all sickchildren under two years have a feeding assessmenteven if they have a normal Z-score (WHO, 1998).The standard deviation unit or Z-score is thesimplest way of making comparisons to thereference population. The Z-score is defined asthe difference between the value for an individualand the median value of the reference populationin the same age or weight, divided by the standarddeviation of the reference population. The medianis the value at exactly the mid-point between thelargest and smallest.How to Measure ItThe measurement of this indicator is based onobservations of the performance of health workerswith regard to the assessment of feeding practicesfor sick children who are brought to a healthfacility. The observer records on a questionnaireor checklist whether the health worker asked aboutor the care taker reported on whether the child isbreastfed, whether the child is taking any otherfoods or fluids, and whether feeding practices hadchanged during the illness. Routine supervisoryvisits can also yield data for measuring thisindicator.Exit interviews are a second option for collectingdata to measure this indicator. The caretaker isasked whether the health worker had asked him/her questions about breastfeeding, complementaryfeeding, and changes in the child’s feeding duringthe current illness.Only a sick child less than two years old who hasa “yes” answer on all three aspects of the assessmentof feeding practices is included in the numerator.The denominator is the total number of sickchildren under the age of two years who were seenby the health worker.Strengths and LimitationsMethodologically, this indicator is relatively easyto construct. The indicator could be aggregatedacross all health workers in a facility to calculatean average child feeding assessment score for agiven health facility. This is useful for identifyingfacilities that are performing below standards withIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills223

egard to the assessment of feeding practices. Alow score on the indicator is a fairly soundindication of the need for refresher training ortargeted supervision.There are, however, difficulties in interpretingchanges in this indicator. Lack of change in theindicator could mean deterioration in one or moreaspects of child feeding assessment counteractedby improvements in other aspects, or could merelyindicate a general lack of performanceimprovement in this component of sick childassessment. For purposes of planning andmonitoring, it may be useful to monitor changesin the individual components of the indicator thatare of most interest to program managers. Thus,programs may want to calculate a separateindicator for each task associated with theassessment of child feeding in order to identifywhich tasks are not routinely performed.Sample QuestionsThe following are sample questions from the SPA(dated March 2004):o Record whether a provider asked about orperformed other assessments of the child’shealth by doing any of the following:(1) Offer the child something to drink or askthe mother to put the child to the breast(to find out whether the child can drink);(2) Ask about normal breastfeeding practiceswhen the child is not ill; and(3) Ask about feeding or breastfeedingpractices for the child during this illness.224Chapter 6

CHILDNEEDINEEDING AN ORALANTIBIONTIBIOTICTIC AND/ORANTIMALNTIMALARIALARIALIS PRESCRIBEDTHE DRUGUG(S) ) CORRECTLORRECTLYPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children who do not needurgent referral, who need an oral antibiotic and/or an antimalarial who are prescribed the drug(s)correctly.Numerator: Number of sick children withvalidated classifications, who do not needurgent referral, who need an oral antibioticand/or an antimalarial (pneumonia, and/ordysentery, and/or malaria, and/or acute earinfection, and/or anemia in high malaria riskareas) who are prescribed the drug(s) correctly,including dose, number of times per day, andnumber of days.Denominator: Number of sick children withvalidated classifications who do not needurgent referral and who need an oral antibioticand/or an antimalarial.Validated Classifications: The validatorclassification using the IMCI standard protocolis considered to be the gold standard and as closeto the actual diagnosis as is possible to get in theoutpatient setting.Measurement ToolsHealth Facility Assessment (HFA); supervisionchecklistsWhat It MeasuresThis indicator assesses the ability of the healthworker to provide correct treatment, given correctidentification of common childhood diseases.How to Measure ItThe indicator is a composite measure and is restrictedto children with validated classifications. The dataare collected through direct observation of sick childconsultations during a health facility survey or routinesupervisory visits. The following pieces ofinformation are needed to calculate the denominator:(1) Whether or not a child needs urgent referral.• Children who need urgent referral areexcluded from the both the numerator andthe denominator of this indicator.(2) Whether or not a child needs an oral antibioticand/or an antimalarial.• Oral antibiotics and/or antimalarials shouldbe given for pneumonia, and/or dysentery,and/or malaria, and/or acute ear infection,and/or anemia in high malaria risk areas.Children are included in the numerator if each ofthe following elements is prescribed correctly bythe health worker:(i) How many tablets or capsules or spoonfuls totake each time(ii) How many times a day to give the medication(iii)How many days to continue treatmentIf a child needs both an antibiotic and an antimalarial,the health worker must prescribe each drug correctlyfor the child to be included in the numerator. TheIMCI chart shows how many days and how manytimes each day to give antibiotics and antimalarialsdepending on the child’s weight or age. Note thatoral antimalarials vary by country. In high malariarisk areas, the health worker should prescribe anappropriate antimalarial. This should be theantimalarial chosen by the National Drug PolicyProgram, which assesses the rate of resistance toIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills225

different anti-microbials and decides on the drugsmost appropriate for the treatment of uncomplicatedmalaria.Strengths and LimitationsOne of the limitations of this indicator is that it isnot straightforward to calculate. It addresses multipledimensions of the quality of sick child assessmentand has many components. First, the health workermust prescribe each drug correctly in terms of allthree of the following elements: how many tabletsor capsules or spoonfuls to take each time; how manytimes a day to give the medication; and how manydays to continue treatment. The age or weight ofthe child should also be considered in determiningthe correct dose of antibiotic or antimalarial. If achild has multiple classifications, each classificationand prescribed treatment should be considered inthe calculations.Given the number of components that are includedin the calculations, it would be difficult to interpretchange in the indicator. The indicator requires thatall required screening and assessment tasks have beenperformed correctly and, therefore, may notdemonstrate change until these other preliminarytasks have been mastered. Second, correctclassification may not be possible if the health workerdoes not have some item of essential equipment, suchas a timing device for counting respiratory rate or ascale for measuring weight.Note that this indicator cannot be calculated fromthe March 2004 version of the SPA instruments.Although the SPA questionnaire for the observationof sick child consultation requires that the observernotes the diagnosis and treatment prescribed for theillness, the observer is not required to record the dose,number of times per day, or number of days for anymedication that is prescribed. Earlier questions onlynoted whether the provider explained how toadminister oral treatments, asked the caretaker torepeat the instructions for the medications, and gavethe first dose of the oral treatment. While thesequestions permit an assessment of whether onecomponent of counseling (i.e. teaching the motheror caretaker how to give oral drugs at home) isperformed, they do not enable a determination ofwhether antibiotics and/or antimalarials wereprescribed correctly.226Chapter 6

SICICK CHILDNOT NEEDINEEDING ANTIBIONTIBIOTICTIC LEAEAVESTHEFACILITCILITY WITHOUT ANTIBIONTIBIOTICTICPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of children who do not need urgentreferral and who do not need an antibiotic for oneor more IMCI classifications who leave the facilitywithout having received or having been prescribedantibiotics.Numerator: Number of children with validatedclassification who do not need urgent referraland do not need an antibiotic for one or moreIMCI classifications (no-pneumonia cough orcold; diarrhea with or without dehydration;persistent diarrhea; malaria; fever that is notlikely to be due to malaria; measles; chronicear infection; anemia or very low weight-forage;no anemia and not very low weight) wholeave the facility without receiving antibioticsor a prescription for antibiotics for thosevalidated classifications.Denominator: Number of sick children withvalidated classifications who do not needurgent referral and who do not need anantibiotic for one or more IMCIclassifications.Measurement ToolsHFA; SPA; supervision checklistWhat It MeasuresThis indicator measures the ability of healthworkers to correctly differentiate betweenconditions needing antibiotics and those that donot. Unnecessary treatment with antibiotics is notonly ineffective, but it presents an unnecessary riskto the child through possible side effects andallergic reactions. Furthermore, the widespreadunnecessary use of antibiotics encourages thedevelopment of bacteria that are resistant to thedrugs.How to Measure ItThe measurement of this indicator relies onobservations of sick child consultations and acomparison of the classification and treatment ofsick children with the IMCI “gold standard.” Avalidator reexamines each sick child, or a sampleof cases seen by the health worker in order todetermine whether the correct diagnosis andtreatment were given.The denominator consists of children withvalidated classifications who do not need urgentreferral and do not need an antibiotic for one ormore IMCI classifications. The IMCIclassifications for which antibiotics are not neededare the following: no pneumonia cough or cold;diarrhea, with or without dehydration; persistentdiarrhea; malaria; fever that is not likely to be dueto malaria; measles; chronic ear infection; anemiaor very low weight-for-age; no anemia and notvery low weight. The numerator is the number ofchildren meeting these criteria who leave thefacility without receiving or being prescribed anantibiotic.Strengths and LimitationsThis indicator can be used to assess patterns ofover-treatment with antibiotic. As with allcomposite indicators a lack of change in theindicator could signify that improvements in someareas may mask deteriorations in others. Programmanagers may want to see scores reportedseparately by type of classification in order toidentify areas of weakness and improve trainingprograms.Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills227

CARETARETAKERAKER OF SICICK CHILDIS ADVISEDTO GIVEEXTRAFLUIDSAND CONTINUEFEEDINEEDINGPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children with validatedclassifications who do not need urgent referralwhose caretakers are advised to give extra fluidand continue feeding.Numerator: Number of sick children withvalidated classifications, who do not needurgent referral, whose caretakers are advisedto give extra fluid and continue feeding.Denominator: Number of sick children withvalidated classifications, who do not needurgent referral.Validated classification: All sick children need tobe classified correctly by the health worker. Thevalidation of the health worker’s classificationneeds to done at the health center by an authorizedperson (basically a doctor or clinical supervisor)who reexamines a sample of cases seen by thehealth worker to see if a correct classification wasmade.Measurement ToolsSPA; supervisory checklistWhat It MeasuresThis indicator measures the extent to which healthworkers are complying with standards forcounseling a mother or caretaker of a sick childon the need to continue feeding and increase thechild’s fluid intake at home. Sick children needto increase their fluid intake during and afterillness to avoid dehydration. This is especiallyimportant for diarrheal illness but is also true forother illnesses that may make children less likelyto drink. Continued feeding (includingbreastfeeding) during illness shortens the durationof the illness episode and reduces the risks ofdehydration and growth faltering.How to Measure ItThe information required for this indicator canbe obtained during a health facility survey throughdirect observations of sick child consultations orexit interviews of caretakers. A health worker isincluded in the numerator if he or she scorespositively on both advice about the need tocontinue feeding during illness and advice aboutthe need to increase the sick child’s fluid intake.The denominator comprises sick children withvalidated classifications who do not need urgentreferral.There are two groups of interest: (1) sick childrenaged zero through five months and (2) sickchildren aged six to 23 months. The former shouldbe counseled for continued and increased ondemandbreastfeeding during illness (no foods orliquids). The latter should be counseled forcontinued and increased breastfeeding on-demand(greater frequency and longer), age-appropriatefeeding recommendations, and general increase offluids (breastfeeding and other).Strengths and LimitationsThe indicator does not measure whether themother or caretaker complies with the advice orwhether the advice on continued feeding specifiesthe types of food and frequency of feedingrecommended for the child’s age. As with manyother indicators in this chapter, observation biasis of concern if the data are collected through directobservations of sick child consultations. Althoughexit interviews of caretakers may be a more cost-Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills229

effective method of data collection, some degreeof misreporting of the content of counseling couldoccur. If no changes are observed in this indicatorover time, it may be useful to report separately thetwo elements of this indicator: continued feedingand increased fluid intake during illness.Sample QuestionsSample questions from the SPA (dated March2004) are the following:Option 1: Observation of sick child consultationo Record whether a provider did any of thefollowing when counseling the caretaker:(1) Provide general information about feedingor breastfeeding the child even when not sick• Yes/No/Don’t know/NA(2) Tell the caretaker to give extra fluids tothe child during this sickness• Yes/No/Don’t know/NA(3) Tell the caretaker to continue feeding thechild during this sickness• Yes/No/Don’t know/NAOption 2: Exit interviewo What did the provider tell you about feeding[NAME] during this illness?• Give less than usual• Give same as usual• Give more than usual• Give nothing/not feed• Didn’t discuss• Don’t knowo What did the provider tell you about givingfluids (or breastmilk, if the child is breastfed)to [NAME] during this illness?• Give less than usual• Give same as usual• Give more than usual• Give nothing/not feed• Didn’t discuss• Don’t know230Chapter 6

CHILDNEEDINEEDING VACCINCCINATIONSLEAEAVESFACILITCILITY WITHALLNEEDEDVACCINCCINATIONSPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of children needing vaccinations (basedon vaccination card or history) who leave thehealth facility with all needed vaccinations(according to national immunization schedule).Numerator: Number of children needingvaccinations (based on vaccination card orhistory) who leave the health facility with allneeded vaccinations.Denominator: Number of children seen whoneed vaccinations (based on vaccination cardor history).This indicator refers to children aged two monthsup to five years of age unless otherwise defined ina given country’s national immunization policy.Measurement ToolsSPA; HFAWhat It MeasuresThis indicator helps to assess the performance ofa health worker in incorporating immunizationservices into the management of sick children andthe extent to which “missed opportunity” forvaccination is reduced at a particular health facility.All infants and children should have theirvaccination status checked at every facility visit.Children who are due a vaccination should receiveit on the same day (or be referred to the nextvaccination session). Thus, the indicator capturesnot only a health worker’s assessment of a child’svaccination status and the administration ofneeded vaccines, but also the ability of a healthfacility to maintain an adequate stock of vaccines.How to Measure ItThe data requirements for measuring thisindicator are: (1) the number of children due atleast one vaccination that received all vaccinationsdue and (2) the number of children due at leastone vaccination. In health facility surveys, the dataare collected through observations of sickconsultations and caretaker exit interviews and arelimited, therefore, to sick children under five.During routine supervisory visits, both sickconsultations and well baby visits can be observedwhen collecting data to measure this indicator.Doses of vaccines given will be entered in thehealth facility record for the child, which can alsobe used to verify whether the child left the facilitywith needed vaccinations.The child’s vaccination status is determined byeither the vaccination card or history. In countrieswhere different vaccinations are given to childrenin different locations (for example, DTP in thethigh, measles in the arm), health providers mayuse this information to help the caretaker recallwhich vaccines have been given to the child. Theobserver notes on the questionnaire or checklistwhether the child was due for immunization andif so, whether the provider immunized the child,referred the child for an immunization, or tookno action.During an exit interview, the interviewer asks thecaretaker to see the child’s vaccination card andnotes whether the child has ever received anyvaccinations (Polio-0; BCG; Polio-1; Polio-2;Polio-3; DTP-1; DTP-2; DTP-3; Measles) andthe date on which each vaccination was received.The interviewer also notes whether the child’svaccination record shows that the child wasvaccinated on the day of the health facility visit.It must be noted that the March 2004 version ofthe SPA questionnaire does not collect data onIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills231

the type of vaccinations received for childrenwhose caretakers could not produce a vaccinationcard during the exit interview. Therefore, dataderived from the SPA exit interview are restrictedto children whose vaccination cards were seen bythe interviewer.If health facilities are sufficiently equipped withvaccines, and health care providers trained toperform vaccination according to the nationalimmunization schedule, the percentage shouldapproach 100%.Strengths and LimitationsThis indicator permits an assessment of the extentto which policies regarding the immunization ofsick children are observed. Although the indicatorfocuses on children under age five years, it can becalculated separately for infants under age 12months – the most important age group that needstimely protection. Measuring this indicator forindividual health facilities permits theidentification of facilities where sick children arenot routinely immunized.For program purposes, it might be important toknow whether children in need of vaccination donot receive all vaccinations due on the day of thehealth facility visit because of poor health workerassessment of a child’s vaccination status or becausevaccines are unavailable at the health facility onthe day of the sick child visit or because healthworkers believe erroneously that illness andmalnutrition are a contraindication toimmunization. It is recommended, therefore, thatthe indicator be interpreted in the light of theavailability of the four recommended antigens(BCG, polio, DTP, and measles) at the facility onthe day of the visit and that qualitative researchbe conducted to determine the reason(s) for lowperformance on this indicator. It must berecognized that some lower level health facilitiesmay not vaccinate on a daily basis as this wouldlead to huge amounts of wastage in the case offreeze-dried vaccines like BCG, which comes onlyin 20-dose vials and must be discarded six hoursafter opening.232Chapter 6

ORSCARETARETAKERAKER OF CHILDWHOIS PRESCRIBEDORS AND/ORORALANTIBIONTIBIOTICTIC AND/ORANTIMALNTIMALARIALARIAL KNOWS HOWTO GIVETHE TREAREATMENTPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children prescribed oralrehydration salt (ORS), and/or an oral antibioticand/or an antimalarial whose caretakers candescribe correctly how to give the treatment.Numerator: Number of sick children prescribedORS, and/or an oral antibiotic, and/or an oralantimalarial whose caretakers can describehow to give the correct treatment, includingthe amount, number of times per day, andnumber of days.Denominator: Number of sick childrenprescribed ORS and/or an antibiotic and/oran antimalarial.Measurement ToolsHFAWhat It MeasuresThis indicator measures caretaker knowledge ofhow to give treatment correctly, which is anessential requirement for correct home treatment.The indicator reflects the success of a healthworker’s efforts to counsel the caretaker of a sickchild on how to give oral drugs at home and servesas a proxy measure for correct treatment at home.How to Measure ItThe data required for measuring this indicator arecollected through exit interviews. The interviewerasks the caretaker of a sick child to see themedicines/prescriptions received that day from thehealth center. For each medicine given orprescribed, the caretaker is asked how to give thetreatment. The questionnaire typically includes atable with the names of all the medications(antibiotics, antimalarials, ORS) that are beinggiven at that particular health center and spacesfor recording the caretaker’s responses on howmuch medicine to give to the child each time(number of spoonfuls/tablet), how many times perday, for how many days, and the need to completethe entire course of medication. To helpcaretakers/mothers recall the dosing, theinterviewer may show them samples of the tabletsor capsules that they have been prescribed and askthem to demonstrate how they will give themedication at home.To be classified as “describing correctly how to givethe treatment” the child’s mother or caretaker mustdescribe each element correctly (how muchmedicine to give the child each time, how manytimes per day, how many days). If a child receivedmore than one medication or prescription, thecaretaker is included in the numerator only if heor she can describe correctly how to administereach medication. Note that this indicator cannotbe obtained from the March 2004 version of theSPA questionnaire.Strengths and LimitationsThe data required for this indicator are simple tocollect and useful where systematic efforts aremade at a health facility to counsel all caretakerson how to administer oral treatments correctly.Teaching a mother or caretaker how to give oraldrugs requires a series of simple steps including:(1) telling the mother and caretaker what thetreatment is and why it should be given; (2)demonstrating how to measure a dose; (3)describing the treatment steps; (4) watching themother or caretaker practice measuring a dose; (5)asking the mother or caretaker to give the dose tothe child; (6) explaining carefully how and howIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills233

often to do the treatment at home; (7) explainingthat ALL oral drug tablets or syrups must be usedto finish the course of treatment, even if the childgets better; and (8) checking the mother’s orcaretaker’s understanding. If a mother or caretakerdoes not accurately describe how to administer anoral drug at home, it is difficult to tell from theindicator which steps of the counseling processwere not adequately performed.234Chapter 6

SICICK CHILDNEEDINEEDING REFERRALIS REFERREDPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children needing referral whoare referred by the health workers.Numerator: Number of sick children with avalidated classification of severe diseaseneeding referral (one or more danger signs,severe pneumonia or very severe disease, and/or severe dehydration with any other severeclassification, and/or severe persistent diarrhea,and/or very severe febrile disease, and/or severecomplicated measles, and/or mastoiditis, and/or severe malnutrition or severe anemia) whowere referred by the health workers.Denominator: Number of sick children with avalidated classification of severe diseaseneeding referral.Measurement ToolsSPA; HFAWhat It MeasuresThis indicator measures both the health worker’sability to appropriately assess clinical conditionsneeding referral and appropriately refer.classification; (4) severe persistent diarrhea; (5)very severe febrile disease; (6) severe complicatedmeasles; (7) mastoidosis; (8) severe malnutrition;and (9) severe anemia.In order to have a valid appraisal of the healthworker’s ability to appropriately refer, all childrenneed to be re-assessed by a gold standard examinerin order to validate severe classification. Thisentails a re-examination of the sick childperformed separately by a nurse or physician (orother health worker trained in validationexaminations) in another room without priorknowledge of the results of the health worker’sexamination. The classification is consideredvalidated if the nurse or physician or validatormakes a determination that the child needs to bereferred based on IMCI guidelines.Strengths and LimitationsThis indicator may be difficult to obtain if theproportion of sick children with severe diseaseclassification is low. Additionally, the indicatordoes not measure completion of the referralprocess. Successful referral of severely ill childrendepends on a number of factors including effectivecounseling of the caretaker and the availability oftransportation.How to Measure ItData for calculating this indicator are collectedthrough direct observation of sick childconsultations. For a child to enter thedenominator, he or she should have signs andsymptoms of severe illness which may include oneor more of the following: (1) one or more dangersigns; (2) severe pneumonia or very severe disease;(3) severe dehydration with any other severeIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills235

RefererencesencesBryce J, Victora CG, Habicht JP, Vaughn JP, and Black RE. (2004). The Multi-Country Evaluation ofthe Integrated Management of Childhood Illness Strategy: Lessons for the Evaluation of PublicHealth Interventions. American Journal of Public Health, 94(3):406-415.UNICEF. (1999). The Household and Community Component of IMCI: A Resource Manual on Strategiesand Implementation Steps. Nairobi, Kenya: UNICEF, ESARO. Retrieved June 9, 2005, from http://www.unicef.org/health/files/health_imcimanual.pdf.WHO. (1999). IMCI Indicators, Monitoring and Evaluation. Geneva, Switzerland: World HealthOrganization. WHO/CHS/AH/98.1K. Retrieved June 9, 2005, from http://www.who.int/childadolescent-health/New_Publications/IMCI/WHO_CHS_CAH_98.1/Rev99.K.pdf.WHO. (2001). Model Chapter for Textbooks: IMCI. Geneva, Switzerland: World Health Organization.WHO. (2001). Indicators for IMCI at First-level Facilities and Households. Geneva, Switzerland: WorldHealth Organization. Retrieved June 9, 2005, from http://www.who.int/child-adolescent-health/New_Publications/CHILD_HEALTH/indicators_for_IMCI.htmIntegrated Management of Childhood Illness (IMCI): Improved Health Worker Skills237

Annex 6.1. Additional Indicators for IMCI at the Health Facility LevelChapter 6. Integrated Management of Childhood Illness (IMCI): Improved Health Worker Skills239

7DIARIARRHEAHEA, ACUCUTETERESPSPIRIRATORORYINFECTNFECTIONION, ANDFEVEVERIndicators:• Proportion of households with access to an improved source of drinkingwater• Proportion of households using an improved toilet facility• Proportion of households with access to essential handwashing supplies• Proportion of households storing drinking water safely• Proportion of households treating drinking water effectively• Proportion of households where drinking water has sufficient levels ofresidual chlorine• Proportion of households where the caretaker of the youngest child underfive reported appropriate handwashing behavior• Proportion of households that disposed of the youngest child's feces safelythe last time s/he passed stool• Period prevalence of diarrhea• Child with non-bloody diarrhea treated with antibiotics• Oral rehydration therapy (ORT) use rate• Proportion of children aged 2-59 months with diarrhea in the last twoweeks who were treated with zinc supplements• Proportion of children aged 0-59 months with diarrhea in the last twoweeks who received increased fluids and continued feeding• Period prevalence of acute respiratory infection needing assessment• Care · seeking for ARI in children 0-59 months of age• Period prevalence of history of fever• Child sleeps under an insecticide-treated net• Child with fever receives appropriate antimalarial treatment• Caretaker knows at least two signs for seeking care immediately• Number of malaria cases among under-fives• Malaria death rate among under-fives

HAPTER 7. D, ACHAPTER7. DIARRHEAIARRHEA, , ACUTERESPIRAESPIRATORORYINFECTIONNFECTION, AND FEVERDiarrhea, acute respiratory infections (ARI),and malaria account for about 40% of allchildhood deaths worldwide. Most childhooddeaths due to these three diseases can be easilyprevented and treated at home or in healthfacilities. For example, diarrhea can be preventedby good hygiene and sanitary practices. When achild with diarrhea becomes dehydrated, rapid andappropriate treatment is necessary both at homeand in the health facility. Malaria can be preventedby the use of insecticide-treated nets. Once thechild has malaria, rapid and appropriate care isessential. Much less is known about how toprevent respiratory illnesses in children but oncea child has a serious respiratory illness, s/he needsappropriate care by a trained health provider.WHO has developed frameworks along withindicators for monitoring the progress andevaluating the outcomes and impacts ofinterventions that address these three commonchildhood conditions. Beginning in the mid-1980s, vertical programs for the control ofdiarrheal diseases (CDD) and ARI establishedindicators for measuring their success with specificinterventions such as the use of oral rehydrationsalt (ORS) and for evaluating mortality impact.With the movement towards a more holisticapproach to child health, diarrhea, and pneumoniamorbidity and mortality (as well as other keyhealth concerns in children, includingmalnutrition and malaria) came to be addressedwithin the IMCI framework. The frameworkcombines improvements in the case managementskills of health workers and in the health systemsrequired to deliver quality care, with improvementsin household and community practices for childsurvival, growth, and development.The household and community component ofIMCI aims to prevent common childhoodillnesses; improve the household and communityresponse to childhood illness and the quality ofcare provided at home; improve appropriate andtimely care seeking behavior when children needadditional assistance outside the home; increasecompliance to recommended treatment and advicefrom trained care providers; and promote asupportive and enabling environment at thehousehold and community level for children’ssurvival, growth, and development. Thehousehold practices identified and agreed uponby UNICEF/ESAR, WHO/AFRO, and otherpartners fall into four main categories: growthpromotion and development; disease prevention;home management; and care seeking andcompliance to treatment and advice (UNICEF,1999). As depicted in Table 7.1, the list ofproposed indicators for the household andcommunity component of IMCI include somethat are of direct relevance to diarrhea, ARI, andfever.Three hygiene interventions – handwashing withsoap; point-of-use (household, health facility,school, etc.) water treatment and safe storage andhandling; and safe disposal of human feces – havebeen shown consistently to reduce diarrheal diseasemorbidity. These hygiene interventions are centralto the Hygiene Improvement Framework (HIF)adopted by USAID, UNICEF, The World Bank,and the Water Supply and SanitationCollaborative Council (WSSCC), and haverecently received favorable attention at the WorldSummit on Sustainable Development, the 2003G8 meeting in Evian, and the launching of theHealthy Environments for Children Alliance.“Halving, by 2015, the proportion of peoplewithout sustainable access to safe water and basicsanitation” are two important MDG targets and ahigh priority area for sustainable development.This section draws on water and sanitationChapter 7. Diarrhea, Acute Respiratory Infection, and Fever241

Table 7.1. List of indicators for the household and community component of IMCII. PRIORITY INDICATORSNutritionX Child under 4 months of age is exclusively breastfedX Child aged 6-9 months receives breastmilk and complementary feedingX Child under 2 years of age is low weight for agePreventionX Child 12-23 months of age is vaccinated against measles before 12 months of ageX Child sleeps under an insecticide-treated net (in malaria risk areas)X Child has feces disposed safely (key family practice under community IMCI [C-IMCI])X Child whose caretaker has appropriate handwashing behavior (key family practice under C-IMCI)Home case managementX Sick child is offered increased fluids and continued feedingX Child with fever receives appropriate antimalarial treatment (in malaria-risk areas)Care seekingX Caretaker knows at least two signs for seeking care immediatelyII. SUPPLEMENTAL INDICATORSNutritionX Continued breastfeeding rate of children aged 12-15 monthsX Complementary feeding frequencyX Stunting prevalenceX Wasting prevalenceX Mean weight-for-age z-scoreX Mean height-for-age z-scoreX Mean weight-for-height z-scorePreventionX DPT vaccine coverageX Polio vaccine coverageX Tuberculosis vaccine coverageX Vitamin A supplementationHome case managementX Ownership of mother's card for children under two yearsMorbidityX Prevalence of night-blindnessX Period prevalence of history of feverX Prevalence of malaria parasitemiaX Period prevalence of diarrheaX Period prevalence of acute respiratory infections needing assessmentSource: WHO (1999).242Chapter 7

measures that have been proposed in the GlobalWater Supply and Sanitation Assessment 2000 Report(WHO and UNICEF, 2000) by the UnitedNations Commission on SustainableDevelopment, in the Water and SanitationIndicators Measurement Guide developed by theFood and Nutrition Technical Assistance(FANTA) Project (Billig, Bendahmane, andSwindale, 1999), and in Assessing HygieneImprovement: Guidelines for Household andCommunity Levels by the Environmental HealthProject (EHP, 2004). The water supply, sanitation,and hygiene-related indicators presented in thisguide are based on the hygiene improvementframework (HIF). Hygiene improvement is acomprehensive approach to prevent childhooddiarrhea through a combination of improvingaccess to water and sanitation hardware andhousehold technologies, promoting properhygiene, and strengthening the enablingenvironment to ensure the sustainability of hygieneimprovement activities.Regarding the monitoring and evaluation ofinterventions that address fever, the RBMinitiative has proposed a framework and set ofindicators for monitoring the principal malariainterventions and related efforts to reinforce thehealth sector. The initiative was launched in 1998and has as its main objective to halve the malariaburden by the year 2010. The framework identifiesfive critical areas for monitoring the impact andoutcomes of RBM. These areas include: (1) theimpact on malaria burden (i.e., mortality,morbidity, and economic losses); (2) malariaprevention and disease management including theprevention and control of epidemics; (3) healthsector development; (4) intersectoral linkages; and(5) support and partnerships (Remme, Binka, andNabarro, 2001). A list of proposed indicators formonitoring and evaluating RBM is given for eachof these critical areas in Table 7.2.Drawing on existing frameworks and guidelines,this section of the guide presents five categoriesof indicators for monitoring and evaluatingdiarrheal diseases (DD), ARI, and malariaprevention programs at the population level: (1)prevention; (2) home case management; (3) careseeking; (4) morbidity; and (5) impact. Theindicators are the following:Prevention• Proportion of households with access to animproved source of drinking water• Proportion of households using an improvedtoilet facility• Proportion of households with access toessential handwashing supplies• Proportion of households storing drinkingwater safely• Proportion of households treating drinkingwater effectively• Proportion of households where drinkingwater has sufficient levels of residual chlorine• Proportion of households where the caretakerof the youngest child under five reportedappropriate handwashing behavior• Proportion of households that disposed of theyoungest child's feces safely the last time s/hepassed stool• Child sleeps under an insecticide treated netHome case management• Child with non-bloody diarrhea treated withantibiotics• Oral rehydration therapy (ORT) use rate• Proportion of children aged 2-59 months withdiarrhea in the last two weeks who were treatedwith zinc supplements• Proportion of children aged 0-59 months withdiarrhea in the last two weeks who receivedincreased fluids and continued feeding• Child with fever receives appropriateantimalarial treatmentDiarrhea, Acute Respiratory Infection, and Fever243

Table 7.2. RBM core indicatorsI. IMPACTXXXXXCrude death rate among target groupsMalaria death rate (probable and confirmed cases) among target groups% of probable and confirmed malaria deaths among patients with severe malaria admitted to a health facilityNumber of cases of severe malaria (probable and confirmed) among target groupsAnnual Parasite Incidence (API) among target groups (by region/according to the epidemiological situation)II. MALARIA PREVENTION AND DISEASE MANAGEMENTPreventionX % of countries having introduced pyrethroids for public health use and insecticide-treated materials in the list ofessential drugs and materialsX % of service providers (health personnel, community health worker [CHW]) trained in techniques of treatment ofnets and/or indoor spraying according to national policyX % of households having at least one treated bednetX % of pregnant women who have taken chemoprophylaxis or intermittent drug treatment, according to the nationaldrug policyX % of antenatal clinic staff trained in intermittent preventive antimalarial treatment for pregnant womenPrevention and control of epidemicsX % of countries with epidemic-prone areas/situation having a national preparedness plan of action for early detectionand control of epidemicsX % of epidemics detected within two weeks of onset and properly controlledEarly diagnosis and prompt treatmentX % of health personnel involved in patient care trained in malaria case management and IMCIX % of health facilities able to confirm malaria diagnosis according to national policy (microscopy, rapid test, etc.)X % of patients hospitalized with a diagnosis of severe malaria and receiving correct antimalarial and supportive treatmentaccording to national guidelinesX % of patients with uncomplicated malaria getting correct treatment at health facility and community levels accordingto national guidelines within four hours of onset of symptomsIII. HEALTH SECTOR DEVELOPMENTHealth policyX % of districts with plans of action reflecting national health policyX % of districts using health information for planningX % of countries having a policy of universal coverage for all with a basic package including relevantmalaria control activitiesService deliveryX % of health facilities reporting no disruption of stock of antimalarial drugs, as specified in the national drug policy, formore than one week during the previous three monthsCommunity actionX % of countries having national guidelines for malaria prevention and treatment including training of all the informalhealth providers and recommendations for home treatment of febrile illness/suspected malaria, recognition of themost frequent signs of danger for children, prevention of malaria during pregnancy, and use of insecticide-treated netsX % of villages/communities with at least one CHW trained in management of fever and recognition of severefebrile illnessX % of mothers/caretakers able to recognize signs and symptoms of danger of febrile illness in a child < five years244Chapter 7

Table 7.2. RBM core indicators (continued)IV. INTERSECTORAL LINKAGESXX% of countries with multisectoral and inter-agency partnerships established% of countries having established official linkages, including the elaboration of research agenda of public healthinterest, between research institutions and the Ministry of HealthV. SUPPORT/PARTNERSHIPXXX% of countries with agreed national RBM budget met by donor funding% of countries with functional sentinel sites for surveillance efficacy of 1st and 2nd line antimalarial drugsNumber of antimalarial drugs which have progressed to the level of phase II trialsSource: WHO and UNICEF (2000).Care seeking• Caretaker knows at least two signs for seekingcare immediately• Care seeking for ARI in children 0-59 monthsof ageMorbidity• Period prevalence of diarrhea• Period prevalence of acute respiratoryinfection needing assessment• Period prevalence of history of feverImpact• Number of malaria cases among under-fives• Malaria death rate among under-fivesThe indicators presented above are not meant torepresent the full universe of measures that havebeen proposed for monitoring and evaluating theoutcomes and impact of programs and initiativesdesigned to reduce DD, ARI, and malaria. Theyare, rather, a small number of population-basedmeasures that have been developed through broadconsensus among many partners and disciplines.For some preventive measures, two separateindicators are proposed to maintain comparabilitywith standard indicators that are included in majorhousehold surveys such as the DHS, MICS, andthe WHO World Health Survey (WHS), and toobtain information that is programmatically morerelevant. The DHS, MICS, and WHS are generalpurpose surveys that can only contain a limitednumber of survey questions and are mainly usedfor national and international comparisons.Programs usually need more detailed informationabout specific interventions, which is representedin the additional indicator.The impact indicators presented here are malariaspecific.Other impact indicators pertaining toinfant and child mortality are presented in Chapternine. We also do not also present indicators formonitoring some of the cross-cutting elementsfrom other components of child health programsthat make significant contributions to thereduction in DD, ARI, and malaria morbidity andmortality. These elements include breastfeeding,immunization, and nutrition, which are directlyaddressed in Chapters four and eight.Diarrhea, Acute Respiratory Infection, and Fever245

PROPOROPORTIONOF HOUSEHOLDSWITH ACCESSTO ANIMPRMPROVEDSOUROURCEOF DRINKINRINKING WATERMDG G Indicator and Core e HygieneImprovement IndicatorDefinitionProportion of households with access to animproved source of drinking water.Numerator: Number of households with accessto an improved source of drinking water.Denominator: Total number of householdssurveyed.The term “water source” is used as a synonym forwater distribution or supply point.An “improved water source” is defined as any of thefollowing types of drinking water supply:• Piped water into dwelling• Piped into yard, plot, or apartment building• Public tap/standpipe• Tubewell/borehole• Protected dug well• Protected spring• Rainwater• Bottled water where combined with pipedwater into dwelling or other improved sourceAn “unimproved water source” includes:• Unprotected dug well• Unprotected spring• Surface water (river, dam, lake, pond, stream,canal, and irrigation channels)• Cart with small tank/drum• Tanker-truck• Bottled water where combined with anunimproved source• Any other type of supplyDistinguishing between protected water sources,which are “improved,” and those that areunprotected, which are “not improved,” is a majorchallenge facing household survey participants. Asdefined in this guide, a protected water source isconstructed in a manner that prevents water frombeing contaminated, particularly by surface runoff(i.e., water from rain, snow melt, or irrigation thatflows over the ground). Protected dug wells haveraised well linings or casings and platforms fordiverting spilled water and are covered. Protectedsprings have spring boxes which are built to protectthe spring from runoff and other contamination.Measurement ToolsPopulation-based surveys (e.g., DHS, KPC, andMICS)What It MeasuresThis indicator measures access to an improveddrinking water source, which is used as anapproximation of safe water. To classify water astruly “safe” would require testing for bacteriologicaland chemical contaminants. This is rarely donein household surveys due to cost considerations.To protect a household against frequent episodesof diarrhea, the water source must be both withineasy reach as well as of good quality. An estimated1.1 billion (18%) of the world’s population didnot have access to an improved source for drinkingwater in 2002, according to a WHO/UNICEFJoint Monitoring Programme (JMP) updatepublished in 2004. Access to a water source isalso an indirect indicator of water use. The averageDiarrhea, Acute Respiratory Infection, and Fever247

liters per capita use per day (lcd) can range fromseveral hundred liters with a pipe connection toless than 10 liters when the source is more than akilometer away. Thus, the closer a water source isto a family, the more water it tends to utilize.How to Measure ItThe caretaker of the child or the head of householdis interviewed about the household's main sourceof drinking water. Interviewers should be familiarwith different water supply types: whether piped(into dwelling, into yard/plot, or public); open well(in yard/plot or public); covered well (in yard/plotor public); tubewell or borehole (in yard/plot orpublic); surface (spring, river, stream, pond, lake,or dam); rain water; tanker-truck; bottled water;gravity flow stream; and so forth. When bottledwater is mentioned as the main source for drinkingwater, a second water source for cooking andhygiene needs to be identified by asking a followupquestion. The second water source must beimproved for bottled water to be counted as animproved source of drinking water; or else bottledwater is counted as unimproved. If the householdobtains drinking water from several sources, theinterviewer is required to probe to determine thesource from which the household obtains themajority of its drinking water.There is a possibility that the water source maydiffer according to season. If a household's normal,wet-season source becomes unusable in the dryseason and the household is forced to travel longerdistances to fetch water, the time to reach the dryseason drinking water source must be recorded.The proportion of households with access to animproved water source is calculated by dividingthe numerator by the total number of householdssurveyed. This indicator is a standard DHS/MICS indicator.This indicator should be measured every three tofive years. Target 10 of the U.N. (2000) MDG isto “halve, by 2015, the proportion of peoplewithout sustainable access to safe drinking water.”As water use varies seasonally, baseline and followupsurveys must be conducted during the sameseason if the results are to be comparable and trendanalysis meaningful.Strengths and LimitationsAccess to an improved drinking water source is auseful and practical approximation of water qualityand availability. However, definitions of accessvary, limiting the usefulness of this indicator forcross-national comparisons. For example, theGlobal Water Supply and Assessment Report 2000defines “reasonable access” as “the availability of20 liters per capita per day at a distance no longerthan 1000 meters” of the dwelling. Some surveysrecord the main source of water used at the timeof interview if the source of water varies seasonally.In other surveys (other than DHS and MICS),data on time to get water may not include the timeit takes to get to the source, wait to get water (ifnecessary), and return to the dwelling.As access and volume are concepts that are difficultto measure, sources of water that are thought toprovide safe water are used mostly as a proxy forthis indicator. To improve the comparability ofdata on safe water coverage (Indicator 1), WHO,UNICEF, and USAID under the JMP haveagreed on a standard approach to its measurementfor all major household surveys (i.e., the DHS,MICS and World Health Survey [WHS]). It isrecommended that evaluators note the definitionsthat are used when examining trends in thisindicator.One limitation of this indicator is that it does notdirectly address issues of water quality. It also doesnot reliably measure the quantity of water used.However, it has been shown that the amount oftime to a water source is a surrogate measure forthe quantity of water used. When using thisindicator as a proxy for water use, it is importantto recognize that water use varies seasonally andthat a single interview may not be reliable becausewater needs may vary from day to day, dependingon household activities (e.g., brewing or laundry).Care should be exercised in interpreting trends in248Chapter 7

the indicator as they could be influenced by thetime of the year in which the surveys areconducted, as well by the definitions used foraccess.As mentioned previously, this indicator informsonly about the type of water source, yet even if itis improved, access may be limited because ofexcessive time spent to collect water or intermittentwater availability. For this reason, it isrecommended that the proportion of householdswith access to an improved water source besupplemented by the indicator below. This wouldensure comparability with international targets aswell as enable programs to obtain more relevantinformation.Supplemental IndicatorDefinition: Proportion of households with animproved source of drinking water withinacceptable reach and available daily.Numerator: Number of households with accessto an improved water source, water collectiontime of 30 minutes or less, and no interruptionfor an entire day or longer within the last twoweeks.Denominator: Total number of householdssurveyed.In this formulation, “access” is defined by twoqualities: (1) time to fetch water; and (2)continuity of water supply. “Acceptable reach” means30 minutes or less for the entire process of fetchingwater, which includes going to the water supplypoint, waiting time, filling containers, andreturning to the household. Distance is a lesscommon and less useful measure, because it doesnot adequately represent the effort of bringingwater to the household. The cut-off value of 30minutes is based on research findings. No separatevalues have been established for rural and urbanareas as of this writing. Acceptable access alsorequires that an improved water source is availabledaily, without interruption for an entire day orlonger during the last two weeks.Diarrhea, Acute Respiratory Infection, and FeverThis supplemental indicator is measured using theDHS Environmental Health Module. In additionto asking the caretaker of the child or the head ofhousehold about the household’s main source ofdrinking water, the time required to reach thewater source or the distance from the source isestimated. This is by whatever means oftransportation the household generally uses,whether the members walk or ride bicycles ormotor vehicles. If access is expressed in terms oftime to the source of water, a round trip, in additionto time for queuing and filling containers, isincluded in the calculations. To be counted in thenumerator of the supplemental indicator, ahousehold must have access to at least one of theimproved types of water sources, water collectiontime must be within 30 minutes, and the sourcemust not have been interrupted for an entire dayor longer within the last two weeks (that is, 14days).This supplemental indicator can be disaggregatedinto levels of access based on time for watercollection as follows: no access (more than 30minutes total collection time); basic access(between five to 30 minutes total collection time);intermediate access (on-plot, e.g., single tap inhouse or yard); and optimal access (water is pipedinto the home through multiple taps). These levelsof access can be interpreted in terms of householdwater security (Howard and Bartram, 2003). Thepercentage of households at each level of access isthen calculated by dividing the number ofhouseholds at that level by the total number ofhouseholds in the sample.A major concern surrounding this supplementalindicator is its composite nature. Thesupplemental indicator measures three things:whether a household has access to an improvedsource of water for drinking; whether the time tocollect water from this improved source is 30minutes or less; and whether there was anyinterruption of water supply from the said sourcefor an entire day or longer in the last two weeks.Thus, when examining trends in the supplementalindicator, improvements in some areas of access249

may be masked by deterioration in other areas.For monitoring purposes, it may be more usefulto track the components of Indicator 2 separatelyto identify specific deficiencies and areas forimprovement.Because water collection can pose a significantburden on household members, it is important toknow which family members usually perform thetask of collecting water. Knowing the particularfamily member or members that collect water givesa sense of whether gender and generationaldisparities exist with respect to water collectionresponsibilities.Sample QuestionsPriority questions:*oooooWhat is the main source of drinking waterfor members of your household?If bottled water is the main source, what isthe main source of water used by yourhousehold for other purposes such as cookingor handwashing?How long does it take you to go to yourprincipal water source, get water, and comeback?Who usually goes to this source to fetch thewater for this household?In the last two weeks has the water from thissource been unavailable for at least one wholeday?Optional supplemental questions:*o When this source is not available at any time,what other source of drinking water do youuse for members of this household?* Priority questions are needed for estimating theindicator(s) described. Supplemental questions areoptional and may be useful for calculatingadditional indicators. Not all questions areavailable from the DHS.250 Chapter 7

PROPOROPORTIONOF HOUSEHOLDSUSINSING AN IMPRMPROVEDTOILETFACILITCILITYMDG G Indicator and Core e HygieneImprovement IndicatorDefinitionProportion of households using an improved toiletfacility.Numerator: Number of households that use animproved toilet facility.Denominator: Total number of householdssurveyed.“Use” implies that the household must have accessto a toilet facility at any time for any member andthat a toilet facility should also be located withina convenient distance from the user’s dwelling,bearing in mind night use and use by children andthe elderly. This can be ascertained throughobservation of the accessibility and use of the toiletfacility.“Improved” toilet facilities include:• Flush/pour-flush toilet connected to pipedsewer system• Flush/pour-flush toilet connected to aseptic tank• Flush/pour-flush latrine connected to a pit• Ventilated improved pit (VIP) latrine• Simple pit latrine with slab (slab that canbe cleaned)• Composting toiletNote: A slab must have a smooth surface to becleaned. This can be made of concrete, plastic,clay spread over sticks, or tightly fitting woodenplanks with a smooth upper surface. A slab ofexposed sticks without a smooth surface wouldnot be considered improved.“Unimproved” toilet facilities include:• Flush/pour-flush latrine that emptieselsewhere without connection to a pipedsewage system, septic tank, or pit• Flush/pour-flush latrine with unknowndrainage• Pit latrine without slab/open pit• Bucket latrine (where excreta are manuallyremoved)• Hanging toilet/latrine• Shared facility of the improved type• Open defecation in field or bush, intoplastic bags (‘flying toilets’)• Any other type of defecationMeasurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresHaving easy access to functioning and improvedtoilet facilities is essential for the improvement ofthe hygienic situation of a household. Thisindicator measures access to an improved toiletfacility. Note that all shared facilities are classifiedas unimproved according to the definition adoptedby WHO and UNICEF in 2000. However, thisseems biased against urban areas where sharedfacilities may be the only feasible sanitationsolution.Diarrhea, Acute Respiratory Infection, and Fever251

How to Measure ItThe child's caretaker or household head isinterviewed about the type of toilet facility usedby the household. Households that have a toiletfacility which is not in working order or is notused for other reasons are classified as not using afacility. After the interview, an observation shouldbe carried out to determine if the sanitation facilityis accessible and shows sign of use (e.g., well-wornpath, unobstructed, etc.). For a household to becounted in the numerator, only use of an improvedtoilet facility is required. Note that all sharedfacilities must be classified as unimproved. Target10 of the MDGs is to "halve, by 2015, theproportion of people without sustainable accessto basic sanitation.Strengths and LimitationsThis indicator can be monitored separately forurban and rural areas. However, unless thisindicator is measured in a standard way, it wouldbe of limited use for regional or cross-nationalcomparisons. In some countries, data forcalculating this indicator are routinely collectedat the national and sub-national levels usingpopulation and housing censuses. Administrativeor infrastructure data may also be used. However,administrative data generally refer to existingsanitation facilities, whether or not they are used.Just having a sanitation facility is not sufficient.There must be signs of consistent use by all familymembers if health impact is to be achieved.Household ownership of improved sanitationfacilities is sometimes used as a proxy for thisindicator. Evaluators must note, therefore, thedefinitions that are used when examining trendsin the indicator. To improve the comparability ofdata on sanitation coverage, WHO, UNICEF, andUSAID under the JMP have agreed on a standardapproach to the measurement of this indicator forall major household surveys, such as the DHS,MICS, and WHS.Sanitation may be a sensitive topic in manycultures. Thus, interviewers must be well trainedand as unobtrusive and sensitive as possible. Forgood program design, qualitative research onknowledge of, attitudes towards, and practices ofexcreta disposal is critical. A study should be sureto detail the types of sanitation facilities that arelocally available and not assess the householdagainst types not usually found in the communitybeing examined. However, any locally definedtypes must be classified under one of the improvedor unimproved categories listed earlier; otherwise,findings between different surveys cannot becompared.While use of an improved toilet facility mayrepresent a necessary condition for “sanitary excretadisposal,” the toilet facility may not be hygienic.Thus, it is recommended that this indicator bemeasured in conjunction with a supplementalindicator measuring the proportion of householdsusing an improved, accessible, and hygienic toiletfacility. This would ensure comparability withinternational targets and enable programs to obtainmore relevant information.The definition and metrics of the supplementalindicator are as follows:Supplemental IndicatorDefinition: Proportion of households using animproved, accessible, and hygienic toilet facility.Numerator: Number of households that usean improved, accessible, and hygienic toiletfacility.Denominator: Total number of householdssurveyed.“Accessibility:” Components of access include thefollowing:• Whether the facility is shared;• Proximity to the dwelling; and• Physical access to the toilet facility.252Chapter 7

Specifically, in order for the toilet facility to beclassified as accessible, the following two criteriamust be met:• Toilet facility must be in or attached tothe dwelling or inside the yard; and• Observed signs of use and absence of signsdiscouraging use.The distance to the toilet facility is approximatedby its location in relation to the dwelling or yard.A “hygienic” facility means the absence of visiblefecal matter on exposed surfaces such as the door,floor, seat, walls, etc.This supplemental indicator, which is measuredusing the DHS/Environmental Health Module,takes four conditions into account: (1) the toiletfacility must be one of the improved types; (2) thefacility must be in the dwelling or yard; (3) thefacility must be accessible and show signs of use;and (4) the facility must be hygienic – there mustbe no feces on the floor, seat, or walls. Based onprior research, these seem to be the most criticalconditions. Additional criteria could furtherstrengthen the programmatic relevance ofinformation collected about sanitation, forexample, whether there is a place for hand washingwith water and soap within or next to the facility,the actual distance between the dwelling and toilet,whether the facility has a basic superstructure thatensures privacy, or whether the toilet isfunctioning.After the interview, an observation should becarried out to determine if the sanitation facilityis accessible and shows sign of use (e.g., well-wornpath, unobstructed, etc.), and whether it ishygienic, which is defined as being without visiblefeces on interior toilet surfaces. The primarylimitation of the supplemental indicator is that itis composite in nature. It combines measures ofuse, access, and hygiene conditions and may bedifficult to measure in practice. Disaggregatingthe supplemental indicator into its constituentparts may facilitate the identification of areas ofimprovement, particularly if trends show little orno change in the indicator over time.Sample QuestionsPriority questions:*o What kind of toilet facility do members ofyour household usually use?o Do you share this toilet facility with otherhouseholds?o How many households use this toilet facility?o Where is the toilet facility located?o If flush/pour-flush: is the toilet facility inworking condition?o Where does this toilet facility drain?o Toilet facility observation: observe access tothe facility. Are there obstacles in the path?Are there signs of regular use?o Toilet facility observation: is there human oranimal fecal matter present inside the facilityon the floor, seat, or walls?Supplemental questions:*o Is the drop hole covered by a lid?o Do you have a place where you usually washyour hands, and if so, where is it?* Priority questions are needed for estimating theindicator(s) described. Supplemental questions areoptional and may be useful for calculatingadditional indicators. Not all questions areavailable from the DHS.Diarrhea, Acute Respiratory Infection, and Fever253

PROPOROPORTIONOF HOUSEHOLDSWITH ACCESSTOESSENTIALHANDANDWASHINASHING SUPPLIESCore e Hygiene Improvement IndicatorDefinitionProportion of households with access to essentialhandwashing supplies.Numerator: Number of households that haveaccess to essential handwashing supplies.Denominator: Total number of householdssurveyed.“Access” means that all essential items forhandwashing are either present (visible at the timeof survey) or can be produced within one minute.A special place for hand-washing may not bealways feasible, but ideally one should be locatedin or near the toilet facility or kitchen.“Essential handwashing supplies” include all of thefollowing:(1) Water (stored in separate container than inthe washing device);(2) Soap (or locally available cleansing agent); and(3) Washing device allowing for unassisted handwashing(tap, basin, bucket, sink, or tippy tap).Clean drying materials, such as towels, are notessential because air drying is an acceptablealternative.Disposal of wastewater after handwashing doesnot require specific measures, unlike wastewaterfrom cleaning up children’s stool. However, lettingwastewater from hand washing accumulate inpuddles should be avoided to keep surroundingsdry and to prevent mosquitoes from breeding.Measurement ToolsPopulation-based surveys such as KPC and DHS/Environmental Health (EH) ModuleWhat It MeasuresBasic handwashing is an important element of thecontrol of diarrheal diseases (DD). Handwashingbehavior is strongly influenced by the presence andaccess to water as well as access to essentialhandwashing supplies. To be optimally effective,the handwashing place should be located in closeproximity to the toilet facility so that householdmembers can conveniently wash their hands afterdefecation, or to the place where cooking takesplace so that food preparers can wash their handseasily before preparing food. At a minimum, thehandwashing place should be inside the yard.How to Measure ItData for calculating this indicator are collectedduring a household interview. A question is askedto find out where household members usuallywash their hands. The interviewer then asks toexamine the site and notes whether the sitecontains a water supply (it is desirable but notessential that this is of the improved type, becauseeven handwashing with water unsafe for drinkingcan be effective), a device for containing water andrinsing hands, and a cleansing agent such as soap.These items can either be displayed or broughtout within one minute for the household to qualifyas having access to essential handwashing supplies.To calculate the indicator, divide the number ofhouseholds with access to all essentialhandwashing supplies by the total number ofhouseholds in the sample.Diarrhea, Acute Respiratory Infection, and Fever255

Strengths and LimitationsThe indicator does not measure the use of handwashingsupplies at appropriate times orknowledge of appropriate hand-washingtechniques. Ideally, actual hand-washing practicesshould be observed, but this is often not practicalduring household surveys. Some surveys do notcollect data on all criteria required in the definitionof essential hand-washing supplies. The currentversion of the core questionnaire of the DHS, forexample, does not ask about a handwashing placeand supplies; however the DHS/EH module doesassess all essential criteria plus whether there isclean material for hand-drying, which can be usedwhere relevant to calculate an additional indicator.It is important, therefore, that baseline and followupsurveys use exactly the same methodology tocalculate the indicator so that any measurementbiases would be systematic.ooObservation only: Is there a towel or cloth todry hands? Note if present or brought in oneminute or less.Observation only: Does the towel or clothappear to be clean?* Priority questions are needed for estimating theindicator described. Supplemental questions areoptional and may be useful for calculatingadditional indicators.Sample QuestionsPriority questions:*o Can you show me where you usually wash yourhands and what you use to wash hands?o Observation only: Is there water? Interviewer:turn on tap and/or check container and noteif water is present or brought in one minuteor less.o Observation only: Is there soap or detergentor locally used cleansing agent? Note if presentor brought in one minute or less.o Observation only: Is there a hand-washingdevice such as a tap, basin, bucket, sink, ortippy tap present or brought in one minute orless?Optional supplemental questions:*o Observation only: Does the washing deviceallow unassisted washing and rinsing of bothhands, for example, a tap, basin, bucket, sink,or tippy tap?256Chapter 7

PROPOROPORTIONOF HOUSEHOLDSSTORINORING DRINKINRINKING WATERSAFELAFELYCore e Hygiene Improvement IndicatorDefinitionProportion of households storing drinking watersafely.Numerator: Number of households storingdrinking water safely.Denominator: Total number of householdssurveyed that store drinking water.“Storing drinking water safely:” To store drinkingwater safely, households should:• Cover the container (containers should havea screw-on top/lid or a plate-like cover thatcompletely covers the water storage containerand fits tightly); and• Use a narrow neck container (containersshould have a neck of 3cm or less in diameter);or• Store water in cisterns and/or roof tanks.Measurement ToolsPopulation-based survey such as KPC and DHS/EH ModuleWhat It MeasuresWhile access to safe water is important, it is alsonecessary for a household to store its waterproperly so that it remains safe. That means thatthe water should not be contaminated by exposureto dirt or dust (hence containers should have anarrow neck and be covered).How to Measure ItData for this indicator are collected through ahousehold interview. The interviewer asks howthe household stores its water and then examinesthe container to ascertain if it is narrow-neckedand covered. A household is counted in thenumerator if it meets all criteria for proper waterstorage: that is, use of a covered and narrow-neckcontainer, which limits access, especially bychildren, to drinking water. Cisterns and rooftanks are considered safe because they generallydo not allow individual household members toserve themselves directly and no directobservations of these storage facilities are includedthat could ascertain whether they are safe. Onlyhouseholds that store drinking water are includedin the denominator.Setting a target of 50% of the households storingwater properly is realistic and attainable, but theproportion reached may be lower or higherdepending on where a community is at thebeginning of a safe water storage intervention (i.e.baseline) and the time available to changeknowledge and practices.Strengths and LimitationsThe data required to calculate the indicator aresimple to collect. As in the case of other hygieneimprovement behaviors, practicing safe waterstorage is predicated on knowledge.Additional criteria are not considered in thecalculation of the indicator, but could be includedin an indicator that measures safe watermanagement comprehensively:• Container with tap or spigot• Limited access of children to the drinkingwater (a narrow neck of 3cm or less)Diarrhea, Acute Respiratory Infection, and Fever257

• Use one different clean vessel to transfer wateror pour it out• Regular cleaning of the storage containerIt is useful to distinguish between water storagecontainers from which water is removed bydipping and those from which it is removed bypouring through a spigot (tap) or a spout. Dippingintroduces objects (ladle, cup, dipper, etc.) andoften the hands that hold these objects into storedwater, and can easily negate the benefits of a cover.The importance of a narrow-neck is that ifsufficiently narrow, it will effectively precludedipping as a way of removing water, and willrequire the user to pour water from the containeror remove it through a spigot (tap) at the bottom.Even if a household has easy access to safe water,the members could be at risk if the water storagecontainer is not proper or properly maintained.Sample QuestionsPriority questions:*o How do you store drinking water?o If in containers: may I see the containers,please?o What type of containers are these? (observe)o Are the containers covered? (observe)Optional supplemental questions:*o Observe: where are the water containersplaced?o Who takes water from these containers?o How do you remove water from the drinkingwater container?o What do you use to remove water?o Are the water containers cleaned?o When last were they cleaned?* Priority questions are needed for estimating theindicator(s) described. Supplemental questions areoptional and may be useful for calculatingadditional indicators. Not all questions areavailable from the DHS.258Chapter 7

PROPOROPORTIONOF HOUSEHOLDSTREAREATINTING DRINKINRINKINGWATEREFFECTIVELFFECTIVELYCore e Hygiene Improvement IndicatorDefinitionProportion of households treating drinking watereffectively.Numerator: Number of households treatingdrinking water effectively.Denominator: Total number of householdssurveyed.“Treating drinking water effectively:” Effectivetreatment requires a methodology that removesfecal pathogens from drinking water and theregular (daily) application of the treatment methodto minimize the risk of recontamination.Although some methods such as chlorination andfiltration can protect water for longer than 24hours, it is assumed that most poor householdswill not store large quantities of water, but willcollect water daily. Where cisterns or roof tanksare common and treated chemically, questions mayneed to be adapted to take into account the lessfrequent treatment of large quantities of water.Where water chlorination is promoted (and whereit is feasible), household water should be testedfor residual chlorine (free and total) usinginexpensive test strips.Effective treatment methods are:• Boil for at least one minute• Add bleach/chlorine• Water filter (ceramic, sand, composite)• Solar disinfectionNote: these methods may be used in combinationwith sedimentation or straining through cloth toDiarrhea, Acute Respiratory Infection, and Feverremove solid matter and to increase theeffectiveness of physical or chemical methods.Ineffective methods of treating drinking water, ifnot used in combination with one of the aforementionedeffective treatment methods, are thefollowing:• Let water stand and settle/sedimentation• Strain water through cloth• Any other methodMeasurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresWhile access to a safe source for drinking water isimportant, many households have no such access.In addition, water from many sources that meetthe criteria for “safe improved water sources” mayin fact be contaminated and unsafe to drink. Thismay be due to breaks in water distribution pipes,failures of chlorination systems, cracks in boreholewell casings, contamination of “protected” springsor dug wells, or unsafe rainwater collectionmethods (i.e. from contaminated roofs). Whenproperly practiced, point-of-use or householdtreatment of drinking water is an effective meansof improving water quality and reducingwaterborne disease. Generally, water is treatedafter it is collected and before it is stored (seeindicator on storing drinking water safely on page257).How to Measure ItData for this indicator are collected through ahousehold interview, and may be supplementedby direct observation. In priority questions, theinterviewer asks how the household usually treats259

its drinking water. A household is counted in thenumerator if it meets the criteria for effective watertreatment: that is, habitual use of an approvedmethod. This indicator is a standard DHS/MICSindicator.Optional criteria for an additional indicatorinclude observation of the tools or materials usedfor treatment (i.e., presence of a filter, a bottle ofhypochlorite solution, containers for solardisinfection of water, or a pot used routinely toboil drink water). A simple test of stored waterfor the presence of chlorine can provide usefulobjective data for households that practice pointof-usetreatment with sodium hypochlorite-basedsolutions. Microbiologic evidence of watertreatment would be ideal, where water quality canbe tested, but because of time and cost this mayonly be feasible in special surveys.In a controlled setting, it may be realistic andattainable to set a target of 75% of householdswithout access to an improved drinking watersource treating water effectively. However, theproportion reached may be much lower forprograms operating at a large scale. Realistictargets also depend on where a community is atthe beginning of a point-of-use water treatmentintervention (i.e. baseline) and the time availableto change knowledge and practices. Targets willvary from place to place according to the qualityand integrity of the water sources. Setting a targetfor point-of-use treatment for households withaccess to an improved water source is of a lesserpublic health priority, because for a proportion ofthese households water is safe, and treatment willnot be needed. Water quality tests may be usefulto establish targets, if they are feasible andaffordable; in most situations, knowledge ofwhether community sources are of the improvedor unimproved type will be adequate.An additional indicator could be calculated byincluding only households at highest risk fromwater contamination. In this case, households withpiped water connections that store water becausethe supply may be irregular would be included in260the denominator. Households with piped waterconnections that do not store water would beexcluded from the denominator.Strengths and LimitationsThe data required to calculate this indicator aresimple to collect. As in the case of other hygieneimprovement behaviors, practicing point-of-usewater treatment is predicated on knowledge. Thisindicator is also a composite of the necessary stepsthat a household must take in order to treatdrinking water effectively. For program planningand monitoring, it may be more useful to askdirected and detailed questions relevant to therecommended method of household (also calledpoint-of-use) water treatment. It is useful tocalculate proportions independently forhouseholds that have access to an improved watersource (as defined in the water source indicator)and for those that do not. In addition, programsmay monitor (1) the proportion of households thattreat drinking water effectively according to theirprimary type of water source; and (2) theproportion of households that possess theappropriate tools and materials for point-of-usedrinking water treatment according to theirprimary type of water source.This indicator usually significantly overestimatesthe proportion of households that drink safe waterbecause it measures only whether households“usually” treat their drinking water. Theeffectiveness of water treatment can be measuredmore accurately by including information aboutthe timing of water treatment. A supplementalindicator measuring the percent of householdstreating drinking water effectively in the past 24hours can be calculated, as shown on page 261. Itis proposed that effective household watertreatment should be measured with both indicators(the proportion of households treating drinkingwater effectively and the proportion of householdstreating drinking water effectively in the past 24hours) to ensure comparability with nationalsurveys and to obtain programmatically morerelevant information.Chapter 7

Supplemental IndicatorDefinition: Proportion of households treatingdrinking water effectively in the past 24 hours.Numerator: Number of households treatingdrinking water effectively in the past 24 hours.Denominator: Total number of householdssurveyed.oHow long do you leave your water in solardisinfection bottles in the sunlight?* Priority questions are needed for estimating theindicator(s) described. Supplemental questions areoptional and may be useful for calculatingadditional indicators. Not all questions areavailable from the DHS.The supplemental indicator is also collectedthrough a household interview and by directobservation. The interviewer asks how thehousehold usually treats its drinking water andwhen the last treatment was applied. The past 24hours is approximated by the responses “yesterday”or “today.”Sample QuestionsPriority questions:*o Do you treat your water in any way to make itsafer to drink?o If yes: What do you usually do to the waterto make it safer to drink?o When did you treat your drinking water thelast time using this method?Optional supplemental questions:*o How often does the household treat itsdrinking water?o If water is treated by a method other thanboiling: May I see the product or device? Askthe respondent to show you the tools ormaterials used for treating drinking water andnote whether they are adequate (i.e., is thefilter broken? Is the container of hypochloritesolution empty? Is the pot for boiling availableand of adequate size? Are there a sufficientnumber of containers for solar disinfection tomeet household drinking water needs?).o How often do you clean your filter?o How often do you repurchase hypochloritesolution?Diarrhea, Acute Respiratory Infection, and Fever261

PROPOROPORTIONOF HOUSEHOLDSWHEREDRINKINRINKING WATERHASSUFFICIENTLEVELEVELS OF RESIDUESIDUALCHLHLORINECore e Hygiene Improvement IndicatorDefinitionProportion of households where drinking waterhas sufficient levels of residual chlorine.Numerator: Number of households wheredrinking water has levels of residual chlorinebetween 0.2 and 0.5mg/l or greater.Denominator: Total number of householdssurveyed.“Sufficient levels of residual chlorine” is usuallymeasured as free chlorine. Residual chlorine intreated water should be between 0.2-0.5mg/laccording to WHO water quality guidelines.Higher concentrations are unnecessary and leadto a noticeable chlorine taste at 3mg/l or greater.This indicator should only be measured wherehousehold water chlorination is used or wherewater is treated at the source or in the supplysystems such as piped water networks.Measurement ToolsPopulation-based surveyWhat It MeasuresThis is a direct measure of water quality, based onthe evidence that chlorine in concentrations of0.2mg/l over several hours effectively eliminatescommon diarrhea-causing pathogens. However,it does not remove chemical contaminants anddoes not kill some common protozoa.graduations. Different shades of a color, oftenpurple, correspond to different chlorineconcentrations. The lowest positive reading oftencorresponds to a concentration of 0.2 mg/l or more,which simplifies the test as only positive andnegative readings can be recorded.Strengths and WeaknessesMeasurable levels of chlorine depend on a varietyof factors besides water treatment. Turbid waterand the presence of sediments tend to bindchlorine, thereby lowering the free level, whichmay become undetectable. The time since the lasttreatment, exposure to sunlight, temperature, andpotential of hydrogen (pH) all influence theconcentration of free chlorine and its ability todisinfect water.Sample QuestionsPriority questions:*o Do you treat your water in any way to make itsafer to drink?If households respond positively to thepreceding water treatment question andmention chlorination, or if water chlorinationwithin the water supply system is standardpractice, perform chlorine dipstick tests.• If bleach, chlorine, or tap water, test waterfor free chlorine• If bleach, chlorine, or tap water, test waterfor total chlorine* Not all questions are available from the DHS.How to Measure ItFree and total chlorine concentrations aremeasured with inexpensive test strips such as thoseused to test swimming pool water or some finerDiarrhea, Acute Respiratory Infection, and Fever263

PROPOROPORTIONOF HOUSEHOLDSWHERETHE CARETARETAKERAKER OFTHEHE YOUNOUNGESGEST CHILDUNDERFIVEREPOREPORTEDAPPRPPROPRIAOPRIATEHANDANDWASHINASHING BEHAEHAVIORCore e Hygiene Improvement IndicatorDefinitionProportion of households where the caretaker ofthe youngest child under five years reported usingsoap for washing hands within the past 24 hoursat two or more critical times (after defecation andone of the following four: after changing a youngchild; before preparing food; before eating; andbefore feeding a child).Numerator: Number of households where thecaretaker of the youngest child under fiveyears reported using soap for washing handswithin the past 24 hours at two or more criticaltimes (after defecation and one of thefollowing four: after changing a young child;before preparing food; before eating; andbefore feeding a child).Denominator: Total number of householdswith children under five years surveyed.The above indicator is based on the assumptionthat in each household only one caretaker (i.e.,the caretaker of the youngest child) will beinterviewed. Where programs decide to interviewmore than one caretaker and assess reportedhandwashing behavior by caretakers of all childrenunder five, the indicator should be calculatedseparately for children 0-23 months and children24-59 months. In addition, the indicatordefinition, numerator and denominator should bemodified as follows:Definition: Proportion of caretakers of childrenaged 0-23 months (or some appropriate age rangeunder five years) who report using soap forwashing hands within the past 24 hours at two ormore critical times (after defecation and one ofthe following four: after changing a young child;before preparing food; before eating; and beforefeeding a child).Numerator: Number of caretakers of childrenaged 0-23 months (or some appropriate agerange under five years) who report using soapfor washing hands within the past 24 hours attwo or more critical times (after defecation andone of the following four: after changing ayoung child; before preparing food; beforeeating; and before feeding a child)Denominator: Total number of caretakers ofchildren aged 0-23 months (or someappropriate age range under five years)surveyed“Appropriate handwashing behavior” includes twodimensions: use of soap and critical times forhandwashing.Critical times for handwashing listed by WHO are:• After defecation• After handling a child’s feces/cleaning babies’bottoms/changing a young child• Before food preparation• Before eating• Before feeding a childMeasurement ToolsPopulation-based surveys (KPC and DHS/Environmental Health Module)What It MeasuresEvidence from trials and observational studiesshow that handwashing with soap reduces the riskof diarrheal disease by 30-50% (Curtis andDiarrhea, Acute Respiratory Infection, and Fever265

Cairncross, 2003). This indicator inquires aboutactual behavior, and not knowledge. In manyinstances, the behavior of the actual caretaker ofthe child (which could be the mother, a sibling,other family, or other help with whom the childspends most of his/her time) and that of thehousehold member who prepares food would bemost important. Handwashing with soap at twocritical times, “after defecation” plus anothercritical time, is suggested as a minimum butprograms may chose to set higher targets if morefrequent handwashing seems achievable.Although ash, sand, and mud are mentioned inthe literature as local alternatives, neither theiracceptability as a cleansing agent nor their actualuse on a significant scale has been established. Theuse of soap is commonly promoted for handwashing through, for example, public-privatepartnerships.How to Measure ItIn a household survey, this indicator is measuredby self-reporting of critical times for handwashing;rarely by demonstration of handwashingtechnique. Data on handwashing techniques arecollected by asking whether the caretaker has soap,has used it in the past 24 hours for hand washing,and the occasions during which soap was used forthis purpose. The 24 hour recall period can beapproximated by respondents mentioning “today”or “yesterday.” If only one caretaker is interviewedper household, all households with children underfive years old surveyed are counted in thedenominator, whether or not they have soap.Where other locally appropriate cleansingmaterials are common (see indicator abouthandwashing supplies), this indicator can becalculated only for households that have soap. Ifmore than one caretaker is interviewed perhousehold, all caretakers of children under five arecounted in the denominator.In past household surveys, caretakers werefrequently asked to name the critical times forwashing hands. The question had multipleanswers and measured knowledge. Interviewerswere instructed not to read the answers out loud,but to record only those mentioned spontaneouslyby the caretaker. Unfortunately, these knowledgequestions had little discriminatory power.Therefore, the soap use questions mentioned aboveare now recommended.Social marketing and health extension/educationprograms have shown that considerableimprovement in handwashing behavior can beachieved over time (Bateman et al., 1995;Whiteford et al., 1996). Targets aimed atincreasing appropriate hand washing by 50% overthe baseline are realistic and attainable.Strengths and WeaknessesAppropriate handwashing behavior includes threedimensions: critical times, frequency, andtechnique. However, handwashing frequency andtechnique are difficult and time-consuming toassess. Requesting a handwashing demonstrationand direct observation of the handwashingtechnique would be desirable, but may beunfeasible in most surveys because it requiresextensive training of the observers and is intrusive,time-consuming, and expensive.Handwashing behavior is strongly influenced bythe presence or absence of a convenient source ofwater. Where water is scarce, people may resortincreasingly to using recycled water for handwashing. Where possible, the use of recycled waterfor handwashing should be assessed during theinterview. Since different methods can be used tocollect data on hand washing, it is important thatbaseline and follow-up surveys use exactly thesame methodology to calculate the indicator sothat any measurement biases would be systematic.It is also important to recognize that this indicatoris based partly on self-reported behavior in thepast 24 hours and does not indicate whetherappropriate hand washing at critical times ispracticed routinely. Note that some large-scalesurveys, such as the ICHS, do not collect data onhand washing.266Chapter 7

Sample Questions*o Do you have soap?o Have you used soap today or did you use soapyesterday?o When you used soap today or yesterday, whatdid you use it for? If “for washing my or mychildren’s hands is mentioned,” probe what wasthe occasion, but do not read the answers.* Note that these questions are not available inthe DHS.Diarrhea, Acute Respiratory Infection, and Fever267

PROPOROPORTIONOF HOUSEHOLDSTHAT DISPOSEDOF THEYOUNOUNGESGEST CHILDHILD’S FECESSAFELAFELY THE LASAST TIMES/HE PASSEDSTOOLS/HCore e Hygiene Improvement IndicatorDefinitionProportion of households that disposed of theyoungest child’s feces safely the last time s/hepassed stool.Numerator: Number of households thatdisposed of the youngest child’s feces safelythe last time s/he passed stool.Denominator: Total number of householdssurveyed.Some programs may decide to interview more thanone caretaker per household or assess safe fecesdisposal for more than one child under five yearsold. Assessing feces disposal for more than onechild makes programmatic sense because thesebehaviors may be different for younger and olderchildren.If feces disposal is assessed for more than one childunder five years old, the indicator 1 should bereported for children aged 0-23 months and 24-59 months separately and the definition,numerator, and denominator modified as follows:Definition: Proportion of children aged 0-23months (or some appropriate age range under fiveyears) whose feces were safely disposed of the lasttime they passed stool.Numerator: Number of children aged 0-23months (or some appropriate age range underfive years) whose feces were safely disposed ofthe last time they passed stool.Denominator: Total number of children aged0-23 months (or some appropriate age rangeunder five years) surveyed.Diarrhea, Acute Respiratory Infection, and Fever“Safe feces disposal:” Passing stool directly into atoilet facility, or throwing it into a toilet facilitywhen a potty is used or defecation occurs in theopen are the safest disposal methods. If feces arerinsed from clothing, bedding, or washable diapers,the water should be discarded in a toilet facility,sewer or septic system, or in a covered grey-waterpit. If households use composting or dry latrine,only the feces can be disposed into such a facility.Urine and wastewater from rinsing must not enterthe composting compartment. Disposable diapersshould not be discarded in a toilet facility becausethey will clog drain pipes and do not decomposefully if they contain plastic materials, which couldmake it impossible to empty the pit. Disposal ofthese diapers in the trash that is collected at leastonce per week could be a relatively safe option,but it may not be available in many developingcountries. Therefore, the latter option is notincluded in the calculation of this indicator. Burialof feces seems an unlikely choice for mosthouseholds, because it is difficult to remove fecesroutinely from the environment in this manner.All other options expose the environment to fecalcontamination. This includes open grey water pitsor garbage pits.The following feces disposal methods are “safe:”• Dropped into toilet facility• Rinsed/washed away• Water discarded into toilet facility(except composting toilet)• Water discarded into sink or tub connectedto drainage system (sewer, septic tank, orpit)The following feces disposal methods are “unsafe:”• Rinsed/washed away269

• Water discarded into composting toilet• Water discarded outside (includes opengrey water pits)• Disposed• Into solid waste/trash but no weekly trashcollection• Somewhere in yard (includes garbage pits)• Outside premises• Buried• Did nothing/left it thereMeasurement ToolsPopulation-based surveys (e.g., DHS, ICHS,KPC, MICS)What It MeasuresThis indicator approximates safe feces disposalpractices by all household members. Feces ofyoung children pose a particular challenge, becauseyoung children are the most likely to contaminatethe immediate household environment and theleast likely to use a sanitation facility. The child isthe unit of observation instead of the caretakerwhen surveys include more than one child underfive years old per household.How to Measure ItThe survey questions for this indicator are adaptedfrom a question included in the current household/woman questionnaires of the DHS and MICS.Thus, the indicator is a DHS/MICS standardindicator. Previous DHS questionnaires wordedthe question slightly differently by askingcaretakers how they “usually” dispose of the child’sfeces. In the most recent version of the DHSquestionnaire, the question refers to the “last time”the youngest child passed stools. This questionformulation may reduce recall bias.Strengths and WeaknessesThe age range for the children surveyed can beadapted to meet local needs without impairingcomparability as long as the child’s age is measuredin such a way as to allow analysis for different agegroups. Children rather than caretakers are usedas a unit of measurement if safe disposal is assessedfor more than one child in the age range targeted.If only one child per caretaker is assessed, theresults based on children or caretakers as the unitof analysis will be the same.Harmonizing indicator definitions is a challenge.It is recommended, therefore, that feces disposalbe measured with two indicators to ensurecomparability with national surveys and to obtainprogrammatically more relevant information. Thesecond proposed indicator is the following:Supplemental IndicatorDefinition: Proportion of households that disposedof the youngest child’s feces appropriately the lasttime s/he passed stoolNumerator: Number of households thatdisposed of the youngest child’s fecesappropriately the last time s/he passed stoolDenominator: Total number of householdssurveyedThe difference between “safe” feces disposal and“appropriate” feces disposal is the specific inclusionof the use and disposal of disposable diapers inthe latter. Disposable diapers are becoming anincreasing waste problem in developing countriesand an increasing environmental hazard withoutproper means of disposal.The following feces disposal methods areconsidered “appropriate:”• Dropped into toilet facility• Rinsed/washed away• Water discarded into toilet facility(except composting toilet)• Water discarded into sink or tub connectedto drainage system (sewer, septic tank, or pit)270Chapter 7

• Disposed• Into solid waste/trash with weeklycollection (except composting toilet)Unsafe feces disposal methods are considered“inappropriate.”Where programs assess appropriate feces disposalfor more than one child under five years old, thedefinition, numerator, and denominator of thesupplemental indicator should be modified asfollows:Definition: Proportion of children aged 0-23months (or some appropriate age range under fiveyears) whose feces were appropriately disposed ofthe last time they passed stool.Numerator: Number of children aged 0-23months (or some appropriate age range underfive years) whose feces were appropriatelydisposed of the last time they passed stool.Denominator: Total number of children aged0-23 months (or some appropriate age rangeunder five years) surveyed.Sample Questionso Where did [NAME OF CHILD] defecatethe last time?o The last time [NAME OF YOUNGESTCHILD] passed stools, what was done todispose of the stools? (Child used toilet orlatrine; put/rinsed into toilet or latrine; put/rinsed into drain or ditch; thrown into garbage;buried; left in the open; other; don’t know.)The disposal of feces into a toilet facility shouldbe cross-checked with the question inquiringabout the household’s use of such a facility.Diarrhea, Acute Respiratory Infection, and Fever271

PERIODPREVREVALENALENCEOF DIARRHEASupplemental IMCI Indicator at theHousehold LevelelDefinitionProportion of children aged 0-59 months who haddiarrhea at any time in the two-week period priorto the survey.Numerator: Number of children aged 0-59months who had diarrhea at any time in thetwo-week period prior to the survey.Denominator: Total number of children aged0-59 months surveyed.“Diarrhea” is commonly defined as three or moreloose or watery stools in a 24-hour period, a loosestool being one that would take the shape of thecontainer. Diarrhea that is of 14 or more days induration is defined as “persistent diarrhea.”Programs may focus on a different age range, forexample, 0-35 or 0-23 months, and the indicatordefinition, numerator, and denominator adjustedaccording to the age range of interest.Measurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresThis indicator measures the prevalence of diarrheaamong children under age five years and givessome indication of the importance of diarrhea asa public health problem. Diarrhea is one of theprincipal causes of morbidity and mortality amongchildren in developing countries, accounting for15-20% of all deaths of children under age fiveyears (Kosek, Bern, and Guerrant, 2003).Diarrhea-related deaths are most commonlyDiarrhea, Acute Respiratory Infection, and Fevercaused by dehydration produced by acute waterydiarrhea and acute dehydration. Death can alsobe caused by infection, particularly in children whohave persistent diarrhea (of 14 or more days induration) and malnutrition, in those who haveother infections at the same time (such aspneumonia), or in those who have bloody diarrhea.How to Measure ItIn a population-based survey, mothers of childrenunder five years of age are asked if their childrenhad diarrhea at any time in the two weekspreceding the survey, and whether they still haddiarrhea in the last 24 hours. In some surveys,the duration of that particular episode of diarrheais also collected. A child is said to have persistentdiarrhea if s/he had diarrhea in the last 24 hoursand if s/he had diarrhea in the last two weeks thatlasted for at least 14 days. Some surveys alsoinclude a question on blood in the stool to providean approximation of the percentage of childrenwho had dysentery. The caretaker may also beasked how many times the child had bowelmovements on the worst day of the diarrhea inorder to get an idea of the severity of the diarrhea.Strengths and LimitationsThe indicator is useful for evaluating theeffectiveness of specific public health interventionsaimed at reducing the frequency of childhooddiarrheal disease. It is simple to calculate and canbe used to examine trends in diarrheal disease overtime. Because diarrheal disease prevalence isinfluenced by season, surveys must occur in thesame season if the data are to be comparable overtime.While it is extremely useful for measuring theimportance of diarrhea as a public health problem,the indicator is a reflection of both old and newcases of diarrhea in the population. It does not273

give any indication of how long the diarrhea haslasted and excludes children who may have diedwith symptoms of diarrhea.Programs may be interested in calculating theperiod prevalence of persistent diarrhea. Althoughpersistent diarrhea (diarrhea that occurs for 14 ormore days) accounts for less than 10% of alldiarrhea cases, it is associated with 30-50% ofdiarrhea deaths (Black, 1993). The indicator canbe modified to calculate the prevalence ofpersistent diarrhea by restricting the numeratorto the number of children aged 0-59 months whohad diarrhea lasting for 14 days or more in thetwo-week period prior to the survey.274Chapter 7

ON-B-BCHILDWITH NONON-BLOODOODY DIARRHEATREAREATEDWITHANTIBIONTIBIOTICSTICSDefinitionProportion of children aged 0-59 months withnon-bloody diarrhea in the last two weeks whowere treated with antibiotics.Numerator: Number of children aged 0-59months with non-bloody diarrhea in the lasttwo weeks who were treated with antibiotics.Denominator: Total number of children aged0-59 months with non-bloody diarrhea in thelast two weeks.Measurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresThis indicator is a measure of inappropriatetreatment of childhood diarrhea. Inappropriatetreatment of childhood diarrhea with antibioticsis widespread and a major public health concern.Non-bloody watery diarrhea is often caused byviruses and does not require antibiotics. Using anantibiotic to treat viral diarrhea may contribute topreventable deaths and waste scarce resources ondrugs that may even aggravate diarrhea episodes.Widespread unnecessary use of antibiotics couldalso lead to antibiotic resistance.How to Measure ItIn a population-based survey, data for calculatingthis indicator are collected for each child born inthe past five years. The mother/caregiver is askedwhether the child had diarrhea in the two weekspreceding the survey, and if so, whether there wasblood in the stool. Respondents are then askedwhether the child received anything to treat thisepisode of diarrhea, and if so, what type oftreatment was given. Subsequent questions askwhether advice was sought from someone else onhow to treat this episode of diarrhea and note alltypes of facility and all persons seen. A child entersthe numerator if s/he had non-bloody diarrheaand was treated with antibiotics. The denominatorconsists of all children under five years old whohad non-bloody diarrhea in the last two weeks.Strengths and LimitationsThis indicator has major programmaticimplications. If a high percentage of children withnon-bloody diarrhea are found to be treated withantibiotics, it can be assumed that IEC programstargeting improvements in home management ofdiarrhea have not succeeded. Widespread use ofantibiotics for the treatment of non-bloodydiarrhea also indicates that health care providersare either advocating the treatment and/or arelacking relevant technical knowledge. Thisindicator does not capture, however, whospecifically is providing inappropriate treatmentof childhood diarrhea. Care providers deliveringinappropriate treatment could be doctors, nurses,clinical staff, pharmacists, or village health workers.This issue is better examined through observationsof sick child consultations.Sample Questionso Has [NAME] had diarrhea in the last twoweeks?o If yes: was there any blood in the stools?o Was anything given to treat the diarrhea?o What was given to treat the diarrhea?• Fluid from ORS packet• Recommended home fluid• Antibiotic pill or syrup• Injection• IV (intravenous)• Herbal remedies/herbal medicines• OtherDiarrhea, Acute Respiratory Infection, and Fever275

(ORT) UORALREHYDRAEHYDRATIONTHERAPY(ORT) USERATEDefinitionProportion of children aged 0-59 months withdiarrhea in the last two weeks who were treatedwith oral rehydration salts and/or recommendedhome fluids.Numerator: Number of children aged 0-59months with diarrhea in the last two weekswho were treated with oral rehydration saltsand/or recommended home fluids.Denominator: Number of children aged 0-59months surveyed who had diarrhea in the lasttwo weeks.The commonly accepted definition of “diarrhea”is three or more loose or watery stools during a24-hour period.“Oral Rehydration Salts” (ORS) refer to a balancedmixture of glucose and electrolytes for use intreating and preventing dehydration, potassiumdepletion, and base deficit due to diarrhea. WhenORS is dissolved in water, the mixture is calledORS solution.A “government-recommended homemade fluid”(RHF) may be a cereal-based mixture or it maybe made from sugar, salt, and water, or it mayinclude soups, or plain water (if nothing else isavailable). Note that a homemade sugar-and-saltsolution is not recommended in some settings dueto the difficulty of getting the quantities right.Measurement ToolsPopulation-based surveys such as DHS, KPC,MICSWhat It MeasuresThis indicator measures program performance incountries where ORS and/or RHF are/is anaccepted part of the diarrheal disease controlprogram. Diarrhea is a principal cause ofmorbidity and mortality among children indeveloping countries. Diarrhea-related deaths aremost commonly caused by dehydration producedby acute watery diarrhea and acute dehydration.The basic principle of home management ofdiarrhea is to prevent dehydration by increasingfluid intake with oral rehydration fluid (includingORS and plain water) or some other governmentrecommendedfluid as soon as the episode ofdiarrhea starts. Increases in the use of ORT havebeen associated with marked falls in the annualnumber of deaths attributable to diarrhea amongchildren under five years in some developingcountries (Victora, Bryce, Fontaine, and Monasch,2000).How to Measure ItIn household surveys, caretakers of children underfive years old with an episode of diarrhea in thelast two weeks are asked whether the child receivedfluid made from the contents of an ORS or a fluidmade from ingredients available in the home andrecommended for use as ORT by the nationaldiarrheal disease control program. Theinstructions for which ingredients to use in therecommended home fluid may vary from countryto country.Strengths and LimitationsThis indicator is easy to measure. It assumescaretaker and community awareness of ORT. Theuse of a two-week reference period to ascertainthe occurrence and treatment of diarrhea decreasesproblems of recall. However, the indicator doesnot capture timely treatment of diarrhea, that is,Diarrhea, Acute Respiratory Infection, and Fever277

whether ORT was provided as soon as the episodeof diarrhea started. The indicator does not alsomeasure whether ORT was prepared appropriately(electrolyte concentration in the case of ORS) orwhether it was administered correctly (in sufficientvolume) to prevent dehydration. It also does nottake into account the severity of illness.The study of time trends in ORT use has beenlimited by the use of four different definitions ofORT in the 1990s. In the early 1980s, theindicator of choice was the proportion of childrenunder five years old with diarrhea in the last twoweeks who were treated with ORS. By 1988, thisdefinition was expanded to include the proportiontreated with ORS and/or RHF. By 1991, ORTwas redefined as increased fluid intake and by1993, continued feeding was included as part ofthe indicator. As demonstrated by Victora et al.(2000), these definitional changes led tofluctuations in ORT coverage over time. Wheninterpreting trends in ORT use, the successivechanges in its definition must be taken intoaccount.Sample Questionso Has [NAME] had diarrhea in the last twoweeks?o Was there any blood in the stools?o Did you seek advice or treatment for thediarrhea from any source? If yes: where didyou seek advice or treatment?o How many days after the diarrhea began didyou first seek advice or treatment for[NAME]?o Does [NAME] still have diarrhea?o Was s/he given any of the following to drinkat any time since s/he started having thediarrhea?(a) A fluid made from a special packet called[LOCAL NAME FOR ORS PACKET]?(b) A pre-packaged ORS liquid?(c) A government-recommended homemadefluid?278Chapter 7

2-59 MPROPOROPORTIONOF CHILDRENAGED2-59 MONTHSWITHDIARRHEAIN THE LASAST TWO WEEKEEKS WHOWERETREAREATEDWITH ZININC SUPPLEMENTUPPLEMENTSDefinitionProportion of children aged 2-59 months withdiarrhea in the last two weeks who were treatedwith zinc supplements.Numerator: Number of children aged 2-59months with diarrhea in the last two weekswho were treated with zinc supplements.Denominator: Total number of children aged2-59 months surveyed with diarrhea in thelast two weeks.The commonly accepted definition of “diarrhea”is three or more loose or watery stools during a24-hour period.Measurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresThis indicator measures the extent to which homeand health facility treatment practices forchildhood diarrhea reflect the recent WHO/UNICEF (2004) joint recommendations thatchildren receive zinc supplements for treatmentof diarrhea. A meta-analysis of acute andpersistent diarrhea showed a 15% reduction in theduration of acute diarrhea and a 24% reduction inthe duration of persistent diarrhea, as well areduction in the severity of diarrhea episodes,among children receiving zinc supplementationcompared with children receiving a placebo.Children receiving zinc supplementation for 10-14 days also showed greater resistance to newepisodes of diarrhea in the two to three monthsfollowing the full treatment and enhanced overallimmune function (Zinc Investigators’Collaborative Group, 1999; 2000). Widespreaduse of ORS and zinc supplementation for diarrheatreatment is estimated to prevent 88% of diarrheadeaths ( Jones, Steketee, Black et al., 2003).How to Measure ItIn household surveys, caretakers of children underfive years old are asked whether the childexperienced an episode of diarrhea in the last twoweeks and if so, what was given to treat thediarrhea. If zinc supplementation is notspontaneously mentioned, the caretaker is askedspecifically whether the child received zincsupplements. The caretaker may be shown asample of zinc tablets or syrup and asked to reportwhether this treatment was given to the child.All children aged 2-59 months with diarrhea inthe last two weeks are counted in the denominator.A child is counted in the numerator only if s/he isaged 2-59 months, had diarrhea in the last twoweeks, and was treated with zinc supplements.Children who received multivitamins with zincin their formulation should not be included in thenumerator if they did not receive therecommended zinc supplements. Althoughmultivitamins with zinc in their formulation mayenhance the overall zinc content of the diet, suchmultivitamins have not been assessed for theirimpact on diarrhea.Strengths and LimitationsThis indicator is easy to measure. It assumesfamily and community understanding of therevised recommendations for managing childhooddiarrhea and that existing zinc formulations areavailable, cost-effective, and easily administeredto infants and children. The use of a two-weekreference period to find the occurrence andtreatment of diarrhea decreases problems of recall.Diarrhea, Acute Respiratory Infection, and Fever279

However, the indicator does not capture therecommended dosage and duration of zincsupplementation for treatment of childhooddiarrhea. WHO and UNICEF (2004) recommenddaily 20mg zinc supplements for 10-14 days forchildren aged 6-59 months and 10mg per day forinfants under six months old with acute diarrhea(i.e., a diarrhea episode lasting less than 14 days).In a household survey, the majority of childrenwith diarrhea in the last two weeks would havehad the illness less than 10 days ago. Thus, itwould not be useful to ask the number of daysthey were given zinc supplements, especially sincethe course of zinc would probably not have startedon the day that the diarrhea began. Additionally,mothers are unlikely to know the number ofmilligrams present in the zinc tablets or syrup usedto treat the child. To assess whether householdpractices for the treatment of childhood diarrheainclude the recommended dosage and duration ofzinc supplementation, a more in-depth studywould be required. Note also that where zincsupplements are promoted in syrup form, theindicator cannot capture the quality of the product,that is, whether the amount of zinc indicated onthe label is present.Sample Questionso Has [NAME] had diarrhea in the last twoweeks?o Was s/he given any of the following todrink at any time since he/she started havingdiarrhea: a fluid from a special packet called[LOCAL NAME FOR ORS PACKET]; apre-packed ORS liquid; a governmentrecommendedhomemade fluid?o Was anything (else) given to treat the diarrhea?o What (else) was given to treat the diarrhea?If zinc is mentioned: How many times was[NAME] given zinc?This indicator should be monitored in conjunctionwith ORT use, increased fluid intake, andincreased fluid and continued feeding. A childwith acute diarrhea should be treated with bothzinc supplements and ORS/RHF (that is, fluidmade from the contents of an ORS packet or fluidmade from ingredients available in the home andrecommended for use as ORT by the nationaldiarrhea control program). Program managersmay want to use a composite indicator measuringthe use of both zinc supplementation and ORS/RHF to treat diarrhea among children in the lasttwo weeks. However, if no change is observed inthis composite indicator over time, it would bedifficult to determine whether this is due to lackof change in both ORT use and zincsupplementation, or whether an improvement inone component of the composite indicator is offsetby deterioration in the other component.280Chapter 7

0-59 MPROPOROPORTIONOF CHILDRENAGED0-59 MONTHSWITHDIARRHEAIN THE LASAST TWO WEEKEEKS WHORECEIVEDINCREASEDFLUIDSAND CONTINUEDFEEDINEEDINGPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of children aged 0-59 months withdiarrhea in the last two weeks who received increasedfluids and continued feeding during the illness.Numerator: Number of children aged 0-59months with diarrhea in the last two weekswho received increased fluids and continuedfeeding during the illness.Denominator: Total number of children aged0-59 months surveyed with diarrhea in thelast two weeks.“Fluids” include oral rehydration salts (ORS),recommended home fluids, and water. Locallyrecommended home fluids may include soups,cereal gruels, yogurt-based drinks, unsweetenedfruit juice, green coconut water, weak tea, plainclean water, or homemade sugar-and-saltsolutions. Note that homemade sugar-and-saltsolution is not recommended in some settings dueto the difficulty of getting the quantities right.Measurement ToolsPopulation-based surveys such as DHS, ICHS,KPC, and MICSWhat It MeasuresThis indicator measures the performance ofprograms aimed at improving home casemanagement of diarrhea. Any illness in childrenis likely to reduce caloric intake and increasechildren’s susceptibility to malnutrition followingeach illness episode. Both increased fluid intakeand continued feeding during illness are importantto reduce this nutritional impact.Diarrhea, Acute Respiratory Infection, and FeverHow to Measure ItData requirements for calculating this indicatorare: (1) the number of children age 0-59 monthswith diarrhea in the past two weeks who receivedincreased fluids and continued feeding; and (2)the total number of children age 0-59 monthssurveyed with diarrhea in the past two weeks.In the DHS and MICS3 questionnaires, data onincreased fluids and continued feeding duringillness are collected only if the child had diarrheain the past two weeks. Caretakers of children whohad diarrhea in the past two weeks are askedwhether the child was given less than usual todrink, about the same amount, or more than usualto drink during the diarrhea. Similarly, questionsabout feeding practices during illness ask whetherthe child was offered less than usual to eat, aboutthe same amount, or more than usual to eat duringthe diarrhea episode, and if less, how much less.In the ICHS conducted by the BASICS II project,fluid intake during illness refers to liquids otherthan breastmilk, and separate questions are askedto find out about breastfeeding practices duringillness. The distinction between breastfeeding andnon-breastfeeding children was made in anattempt to capture feeding practices during illnessamong children who are exclusively breastfed. Ifthe child is currently breastfeeding, the mother isasked whether she changed the frequency ofbreastfeeding while the child was sick. If yes, therespondent is asked whether the frequency ofbreastfeeding was increased, decreased (somewhatless or much less), or if she stopped breastfeedingcompletely during the illness episode. Note thatif ICHS-type questions are used to collect the data,information for both breastfeeding and nonbreastfeedingchildren should be combined whencalculating the numerator.281

Goal 23 of the World Summit for Children is toreduce by 50% the deaths due to diarrhea inchildren under the age of five years and to reducethe diarrhea incidence rate in this age group by25%.Strengths and LimitationsThe “increased fluids” component of the indicatordoes not capture how soon after the start of theillness episode children are offered increased fluids.The timing of the administration of increasedfluids is especially important in diarrhea cases asearly increased fluid intake can prevent manychildren from becoming dehydrated and facilitatecontinued feeding by restoring appetite. Also, theindicator does not measure the nutritional valueof food given to child during illness.The indicator can be disaggregated into anindicator of increased fluid intake during illnessand an indicator of continued feeding duringillness. These separate indicators can be used forthe purpose of tracking the individual componentsof feeding practices during illness, especially if alow value is obtained on the indicator.Sample Questionso Has [NAME] had diarrhea in the last twoweeks?o Now I would like to know how much[NAME] was given to drink during thediarrhea. Was s/he given less than usual todrink, about the same amount, or more thanusual to drink?o IF LESS, PROBE: Was s/he given muchless than usual to drink or somewhat less?o When [NAME] had diarrhea, was s/hegiven less than usual to eat, about the sameamount, more than usual, or nothing to eat?IF LESS, PROBE: Was s/he given muchless than usual to eat or somewhat less?282Chapter 7

PERIODPREVREVALENALENCEOF ACUTERESPIRAESPIRATORORY INFECTIONNEEDINEEDING ASSESSMENTSupplemental Indicator for IMCI atthe Household LevelelDefinitionProportion of children aged 0-59 months reportedto have had a cough and fast and/or difficultbreathing due to a problem in the chest in the lasttwo weeks.Numerator: Number of children aged 0-59months reported to have had a cough and fastand/or difficult breathing due to a problem inthe chest in the last two weeks.Denominator: Total number of children aged0-59 months surveyed.Measurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresThis indicator measures the frequency of existingchildhood cases of acute respiratory infections(ARI) needing assessment in the population andgives some indication of the importance of ARIas a public health problem. Acute lower tractrespiratory infection, primarily pneumonia, is acommon cause of morbidity and death amongchildren under five years of age in developingcountries. Pneumonia is characterized by difficultor rapid breathing and chest-indrawing.How to Measure ItIn household surveys, the identification of ARI isbased on the caretaker’s perceptions of therespiratory symptoms suffered by the child.Mothers are asked whether their children underage five had been ill with a cough in the two weeksprior to the survey, and if so, whether the coughhad been accompanied by short, rapid breaths and/or a problem in the chest and/or a blocked or runnynose. Children are counted in the numerator ifthey meet each of the following three conditions:(1) they had a cough in the last two weeks; (2) thecough was accompanied by fast and/or difficultbreathing; and (3) the cough was accompanied bya problem in the chest with or without a blocked/runny nose.Strengths and LimitationsThe indicator is useful for evaluating theeffectiveness of specific public health interventionsaimed at reducing the frequency of ARI. It issimple to calculate and can be used to examinetrends in ARI over time. However, it should beborne in mind that the data are subjective. Theyare based on the mother’s perception of illness andare not validated by medical examination.It is to be noted that the analysis of time trends inthis indicator and regional or cross-nationalcomparisons of ARI prevalence is limited by theuse of different definitions. In previous versionsof the DHS, ARI was defined as the presence ofcough and fast/difficult breathing in the last twoweeks. In IMCI, the numerator was defined aschildren with difficult breathing, with or withoutcough, in the two weeks preceding the survey. Theargument behind the IMCI definition was thatwhile cough and rapid breathing are symptomsthat are compatible with pneumonia, about 95%of children with pneumonia will have a cough,while a small proportion will not have a cough,but will have difficulty breathing. Currentrecommendations, as are reflected in the definitionpresented here, are to define the numerator aschildren with cough and rapid/or difficultbreathing that is due to a problem in the chest inthe two weeks preceding the survey. Therefore,Diarrhea, Acute Respiratory Infection, and Fever283

successive changes and differences in the definitionof ARI must be taken into account wheninterpreting trends in this indicator or makingregional/cross-national comparisons.Sample Questionso Has [NAME] had an illness with a cough atany time in the last two weeks?o When [NAME] had an illness with a cough,did s/he breathe faster than usual with short,rapid breaths or have difficulty breathing?o When [NAME] had this illness, did s/hehave a problem in the chest or a blocked orrunny nose?284Chapter 7

CARESEEKINEEKING FOROROF AGEOR ARI INHILDREN 0-59 MIN CHILDREN0-59 MONTHSDefinitionProportion of children aged 0-59 months withcough and fast and/or difficult breathing due to aproblem in the chest in the last two weeks whowere taken to an appropriate health provider.Numerator: Number of children aged 0-59months with cough and fast and/or difficultbreathing due to a problem in the chest in thelast two weeks who were taken to anappropriate health provider.Denominator: Total number of children aged0-59 months with cough and fast and/ordifficult breathing due to a problem in thechest in the last two weeks.As defined by UNICEF and WHO, “appropriateproviders” for the diagnosis and treatment of acuterespiratory infection (ARI) include hospitals,health centers, dispensaries, village health workers,mobile/outreach clinics, and private physicians.This definition can be modified according tonational guidelines and local terminology.Measurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresARI, especially pneumonia, is a leading cause ofdeath in children under the age of five indeveloping countries, killing approximately twomillion in 2000. To prevent deaths when childrendevelop ARI, early diagnosis of the condition andappropriate health care are crucial. Goal 24 ofthe World Summit for Children sought areduction by one third in the deaths due to acuterespiratory infection in children under age fiveyears. In the majority of 73 countries where datawere available, more than half of the childrensuffering from ARI were not taken to appropriatehealth care providers. This indicator measures twothings: the knowledge of the caretaker about whenand where to seek care, and actual care-seekingbehavior.How to Measure ItIn a household survey, mothers/caretakers ofchildren under the age of five years are askedwhether the child had a cough and short, rapid,or difficult breathing in the last two weeks andwhether the illness was accompanied by a problemin the chest and/or a blocked or runny nose. Therespondent is also asked whether she sought adviceor treatment for the child’s illness from any sourceand who provided advice or treatment to the child.Some surveys do not collect information oncontact with an appropriate health provider buton contact with health facilities. A child with ARIenters the numerator if s/he was taken to anappropriate provider. Appropriate providers forthe diagnosis and treatment of ARI includehospitals, health centers, dispensaries, villagehealth workers, mobile/outreach clinics, andprivate physicians. If national guidelines do notdefine who is an “appropriate provider” for thediagnosis and treatment of ARI, UNICEF andWHO definitions may be used instead.No global targets exist for this indicator. Targetsetting needs to be based on clear baselineinformation, specific interventions, and realisticestimates based on the likelihood of change(WHO, 1999a).Strengths and LimitationsWhen combined with facility-level information,this indicator enables programs to understand andinterpret observed ARI-related mortality patternsDiarrhea, Acute Respiratory Infection, and Fever285

in the community. For example, if a highproportion of ARI cases seek appropriate care,mortality resulting from ARI should decline.However, there are some methodological concernswith this indicator. First, it is likely to be affectedby caretakers’ perceptions of disease causation andselectivity. If ARI cases among certainsubpopulations are less likely to be reported thanamong other groups, this will affect howrepresentative the indicator is of children with ARIin the general population. Additionally, eventhough ARI is common, a relatively large samplesize is required to measure this indicator in orderto capture the required population for thedenominator, that is, children with ARI in the lasttwo weeks. This has survey cost implications.This indicator assumes that once the child is takento an appropriate provider, s/he will receive qualitycare in terms of appropriate drugs with the rightdosage and proper counseling. Neither the qualityof health facility case management nor caretakercompliance with the treatment are captured by theindicator. The indicator also does not reflect thetimeliness of care seeking. Ideally, children underfive-years of age with recognized danger signs ofmoderate or severe illness should be taken to ahealth facility or trained service provider within48 hours after the commencement of symptoms.Given the focus on early consultation andtreatment for moderate to severely ill children, itis recommended that the indicator be modified asfollows to reflect prompt care-seeking behavior:Proportion of children aged 0-59 months withcough and fast and/or difficult breathing due to aproblem in the chest in the last two weeks whowere taken to an appropriate health providerwithin 48 hours after the illness began.Sample Questionso Has [NAME] had an illness with a cough atany time in the last two weeks?o When [NAME] had an illness with a cough,did s/he breathe faster than usual with short,rapid breaths or have difficulty breathing?ooWhen [NAME] had this illness, did s/hehave a problem in the chest or a blocked orrunny nose?If yes, then:• Did you seek advice or treatment for theillness from any source?– If YES, where did you seek advice ortreatment?• How many days after the illness began didyou first seek advice or treatment for[NAME]?286Chapter 7

PERIODPREVREVALENALENCEOF HISISTORORY OF FEVERSupplemental Indicator for IMCI atthe Household LevelelDefinitionProportion of children aged 0-59 months with areport of fever in the last two weeks.Numerator: Number of children aged 0-59months with a report of fever in the last twoweeks.Denominator: Total number of children aged0-59 months surveyed.Measurement ToolsPopulation-based surveys such as DHS, KPC, andMICSWhat It MeasuresMalaria is one of the three leading causes ofmorbidity and mortality in the developing world.In Africa, malaria accounts for about 20% ofunder-five mortality, excluding neonatal, andconstitutes 10% of the region’s overall diseaseburden mortality (WHO, 2000c). In malariaendemicareas, the major manifestation of malariais fever. Therefore, the prevalence of fever in thesepopulations is used as a proxy for the prevalenceof malaria.Strengths and LimitationsThis indicator is relatively simple to collect andcalculate. It can be used to examine trends overtime in the prevalence of fever, especially as amethod of evaluating the probable impact ofintervention programs in controlled settings. Thedisaggregation of the data by urban and rural area,where possible, can enable an assessment of theextent to which malaria varies across communitiesand help monitor drug resistance.It is difficult, however, to determine the cause ofthe fever. Various infectious diseases areaccompanied by fever, the most common beingmalaria, pneumonia, intestinal infections, measles,and typhoid. Furthermore, the indicator does notallow one to establish whether the fever wasclinically determined or defined based on thecaretaker’s perception that the child’s bodytemperature was above the normal level.Therefore, the severity of the fever cannot beassessed. In some areas, there is seasonal variationof malaria with incidence peaking at certain timesof year. Where malaria is sporadic or seasonal,care should be exercised in interpreting trends inthe indicator as they could be influenced by thetime of the year in which the surveys areconducted. These factors contribute toquestionable validity of the data for establishingthe prevalence of malaria.How to Measure ItIn a household survey, mothers are asked whetherthe child had an episode of fever within the twoweeks prior to the interview. This information isused to calculate the indicator. Data from surveysare generally available every three to five years.Target 8 of the MDGs is to have halted by 2015and begun to reverse the incidence of malaria andother major diseases.Diarrhea, Acute Respiratory Infection, and Fever287

CHILDSLEEPSUNDERAN INSECTICIDENSECTICIDE-TREAREATEDNETPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of children aged 0-59 months whoslept under an insecticide-treated net the previousnight.Numerator: Number of children aged 0-59months who slept under an insecticide-treatednet the previous night.Denominator: Total number of children 0-59months surveyed.An insecticide-treated net (ITN) is a net that hasbeen dipped in an insecticide effective against localmalaria-causing mosquitoes.Measurement ToolsPopulation-based surveys such as DHS, ICHS,KPC, MICSWhat It MeasuresThis indicator is a measure of malaria prevention.The Roll Back Malaria (RBM) initiative,established in 1998 by WHO, the United NationsChildren’s Fund, and the World Bank, identifiesthe use of ITNs as one of the four maininterventions to reduce the burden of malaria inAfrica. Studies conducted in the 1980s and 1990sshowed that the use of ITNs reduced deaths inyoung children by an average of 20% (see, forexample, Alonso et al., 1993; Lengeler et al., 1996).How to Measure ItTo calculate this indicator, two pieces ofinformation are required: the number of childrenaged 0-59 months that slept under an ITN thenight preceding the survey and the total numberof children aged 0-59 months surveyed.Data on ITN-use are usually collected by askingwomen aged 15-49 in households possessingbednets about the use of bednets by all of theirbiological children under five years old. In theICHS, the definition of bednets includes baby nets(nets that are specially designed to fit an infant)and regular mosquito nets that hang over sleepingplaces and can be used to protect family membersagainst mosquitoes while sleeping.Respondents are then asked whether the bednetunder which the child slept on the night precedingthe survey has ever been treated with insecticideto repel mosquitoes or bugs since it was obtained.The next question asks how long ago the bednet(under which the child slept the previous night)was last treated with insecticide. Respondents mayalso be asked how long ago the bednet was boughtor obtained. With the increased availability ofpermanently impregnated nets, additionalquestions must be asked in surveys to obtaininformation on the brand of ITN used by childrenand other household members.Data on coverage of children’s use of insecticidetreatedbednets should be collected about everytwo to three years. Disparities by sex, mother’seducation, and area of residence should be assessed.One of the targets set at the Abuja Summit inApril 2000 was to have 60% of the population atrisk sleeping under ITNs by 2005 (WHO, 2000a).RBM’s goal is to halve the burden of malaria by2010.Diarrhea, Acute Respiratory Infection, and Fever289

Strengths and LimitationsThis indicator allows programs to monitorwidespread use of insecticide-nets to limit humanmosquitocontact. By asking about use of aninsecticide-treated net the night prior to thesurvey, the question aims to reduce recall bias.There are several caveats to this indicator. First,the indicator assumes the insecticide-treated netis maintained and properly used. It has beenshown, however, that the effectiveness of thebednet depends on the condition, age, and the caregiven to the bednet.The prevalence of malaria-carrying mosquitoesvaries seasonally, with a peak during andimmediately following periods of rain. Thus, ITNuse may follow a similar seasonal pattern. Whenevaluating trends in the use of ITNs, attentionshould be paid to the time of the year in whichthe surveys were conducted in order to clarifywhether estimates of ITN use reflect levels duringthe peak or low malaria season.290Chapter 7

CHILDWITH FEVERRECEIVESAPPRPPROPRIAOPRIATEANTIMALNTIMALARIALARIAL TREAREATMENTPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of children aged 0-59 months withfever in the last two weeks who received a locallyrecommended anti-malarial.Numerator: Number of children aged 0-59months with fever in the last two weeks whoreceived a locally recommended anti-malarial.Denominator: Total number of children aged0-59 months surveyed with fever in the lasttwo weeks.The “locally recommended anti-malarial” is theanti-malarial chosen by the National Drug PolicyProgram, which assesses the rate of resistance todifferent anti-microbials and decides on the drugsmost appropriate for the treatment ofuncomplicated malaria.Measurement ToolsPopulation-based surveys such as DHS, KPC,MICSWhat It MeasuresThis indicator measures the extent to whichchildren with fever who are living in high malariarisk areas receive presumptive treatment formalaria. Malaria is one of the three leading causesof morbidity and mortality in the developingworld. In Africa, malaria accounts for about 20%of all childhood mortality below five years of age(WHO, 2000c). Most deaths due to malaria inchildren could be avoided by prompt recognitionand treatment with antimalarial drugs.How to Measure ItThe data requirements for calculating thisindicator are: (1) the number of children aged 0-59 months with fever in the last two weeks whowere treated with a locally recommended antimalarial;and (2) the total number of children aged0-59 months surveyed who had fever in the lasttwo weeks.In DHS surveys conducted in malaria-endemicareas, caretakers of children with fever in the twoweeks preceding the survey are asked what actionswere taken to treat or seek assistance in treatingthe child’s fever. The caretaker is asked if the childwas given any drugs for the fever and if so, whatdrugs were given. All drugs administered to thechild to treat the fever are noted. If a caretakerdoes not know the name of the drug, she is shownsamples of various drugs that are typicallyadministered against fever and asked to identifythe one(s) given to the child. In some surveys,the caretaker is asked how many days after thefever began did the child take antimalarial drugs.One of the targets set at the Abuja Summit inApril 2000 was to have at least 60% of thosesuffering from malaria have prompt access tocorrect, affordable, and appropriate treatmentwithin 24 hours of the onset of symptoms by 2005(WHO, 2000a).Strengths and LimitationsThis indicator is useful for evaluating the impactof information, education and communication(IEC) programs targeted at improving the homemanagement of fever in malaria-endemic areas.However, there are several caveats worthmentioning. Various infectious diseases areaccompanied by fever, the most common beingDiarrhea, Acute Respiratory Infection, and Fever291

malaria, pneumonia, intestinal infections, measles,and typhoid. In high-risk malaria areas, thesensitivity of “any fever” for classification of malariavaries between 98 and 100%. However, thespecificity is low, varying from 6 to 13%. Thiscould lead to over-treatment with antimalarialsin high-risk areas. But classifying all children withfever as having malaria and in need of appropriatetreatment with antimalarials is considered anacceptable strategy because the risk of developingcerebral malaria and other complications ofmalaria is high in infants and young children.Increasing drug resistance requires nationalpolicies to be active in detecting ineffectivetreatments. This means that locally-recommendeddrugs for the treatment of malaria could changeover time to avoid significant increases in malariaassociatedmorbidity and severe disease if drugresistance develops.Sample Questionso Has [NAME] been ill with a fever at any timein the last two weeks?o Did you seek advice or treatment for the illnessfrom any source?o Where did you seek advice or treatment?o How many days after the illness began did youfirst seek advice or treatment for [NAME]?o Is [NAME] still sick with a fever?o At any time during the illness, did [NAME]take any drugs for the illness?• What drugs did [NAME] take?It must be noted that the indicator does not takeinto consideration the duration of the fever, whichcould mean that the child has a more severe illnesssuch as typhoid fever (if the fever lasts for morethan five days). The indicator also does not reflectlevel of malaria risk in the season or consider thepresence of other infections. In a low malariariskseason, children with fever who have evidenceof another infection may not need to be given ananti-malarial. The indicator also does not allowone to assess the timing of the treatment relativeto the onset of fever. It is also important to notethat the indicator only measures whether the childreceived an appropriate antimalarial withoutreference to adequate dosing.In order to assess whether fever is treatedappropriately within 24 hours of the onset ofsymptoms, the indicator can be modified as followsto reflect prompt care-seeking behavior:Proportion of children aged 0-59 months withfever in the last two weeks who received a locallyrecommended anti-malarial within 24 hours ofthe onset of fever.292Chapter 7

CARETARETAKERAKER KNOWS AT LEASEAST TWO SIGNSFORSEEKINEEKING CAREIMMEDIAMMEDIATELTELYPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of caretakers of children aged 0-59months who know at least two signs for seekingimmediate care when their child is sick.Numerator: Number of caretakers of childrenaged 0-59 months who know at least two ofthe following signs for seeking careimmediately: child not able to drink orbreastfeed; child becomes sicker despite homecare; child develops a fever; child has fastbreathing; child has difficult breathing; childhas blood in stool; and child is drinking poorlyDenominator: Number of caretakers ofchildren aged 0-59 months surveyed.“Know” refers to the ability to spontaneously nameat least two signs for seeking care immediatelywhen a child is sick.Measurement ToolsPopulation-based surveys (e.g., ICHS, KPC,MICS); client exit interviewsWhat It MeasuresThe purpose of this indicator is to measurecaretakers’ knowledge and awareness of the dangersigns of childhood illness in order to plan andmonitor the impact of behavior changecommunication (BCC) at the community level.Knowledge of danger signs of childhood illness isan essential first step to appropriate and timelycare seeking, though it is by no means sufficientto ensure it.How to Measure ItData requirements for measuring this indicatorare the number of caretakers of children aged 0-59 months who can spontaneously name at leasttwo of the danger signs of childhood illness listedin the previous column and the total number ofcaretakers of children aged 0-59 months surveyed.In surveys, this information is usually collectedby asking the caretaker to name what signs wouldcause him/her to take a child under the age of fiveyears to a health facility or health workerimmediately. The question usually has multipleanswers indicating various options. Interviewersare instructed not to read the answers out aloud,but to record only those mentioned spontaneouslyby the caretaker.Strengths and LimitationsThe indicator could be used to plan and monitorcommunity BCC efforts. Knowledge of thedanger signs of childhood illness should decreasechild mortality by increasing rates of earlyrecognition of complications and increasing thelikelihood of prompt referral and service use. Theindicator could easily be obtained in householdsurveys and would be responsive to programmaticintervention on the short term. An increase inthe value of the indicator would signifyimprovement in community knowledge of thesewarning signs.However, this indicator only measures knowledge,not behavior. It is widely known that behaviorchange does not always follow increasedknowledge of a given topic. Also, adequateknowledge does not guarantee that a caretaker willrecognize a danger sign in practice. Care seekingis also strongly influenced by cultural beliefs aboutthe etiology of illness. These beliefs may exert amore powerful influence on the mother’s care-Diarrhea, Acute Respiratory Infection, and Fever293

seeking behavior than will knowledge of theappropriate action to take when a sick childexhibits the symptoms outlined above. Therefore,programs that choose to utilize the knowledgeindicator should also investigate barriers (if any)to enacting behavior once knowledge iswidespread.Sample Questionso Sometimes children have severe illnesses andshould be taken immediately to a healthfacility. What types of symptoms would causeyou to take your child to a health facility rightaway?A. FeverB. Seizure/shakingC. Not eating/not able to breastfeed wellD. Drinking poorlyE. Getting sicker/very sick/not getting betterF. Fast breathingG. Difficult breathingH. Chest indrawingI. Blood in stoolJ. Other (specify)K. Don't knowINTERVIEWER: Circle all symptomsmentioned. Do NOT prompt with anysuggestions.294Chapter 7

-FNUMBEROF MALALARIAARIA CASESAMONG UNDERNDER-F-FIVESDefinitionThe number of malaria cases, severe anduncomplicated (probable and confirmed) amongchildren aged 0-59 months in a definedpopulation.WHO standard case definitions are as follows(WHO and UNICEF, 2005):“Uncomplicated malaria:” Any person with feveror fever with headache, back pain, chills, sweats,myalgia, nausea, and vomiting diagnosed clinicallyas malaria.“Probable malaria:” A person with symptoms and/or signs of malaria and who receives antimalarialtreatment.“Confirmed uncomplicated malaria:” Any personwith fever or fever with headache, back pains,chills, sweats, myalgia, nausea and vomiting, andlaboratory confirmation of diagnosis by malariablood film or other diagnostic test for malaria“Severe malaria:” The signs and symptoms ofsevere malaria are the following: cerebral malaria;convulsions; circulatory collapse; abnormalbreathing (pulmonary edema or respiratorydistress syndrome); jaundice; macroscopichemoglobinuria; extreme weakness/prostration(inability to sit without support, or inability to feed,depending on the age of the child), renalimpairment; severe anemia; and hypoglycemia.These manifestations can occur singly, or morecommonly in combination in the same person.The most common and important manifestationsof severe malaria in children are cerebral malaria,severe anemia, respiratory distress, andhypoglycemia (WHO, 2000b; WHO/AFRO,2001).“Probable severe malaria:” A person requiringhospitalization for signs and/or symptoms ofsevere malaria who receives antimalarial treatmentwith laboratory confirmation of diagnosis.“Confirmed severe malaria:” Any personhospitalized with a primary diagnosis of malariaand confirmed by a positive blood smear or otherdiagnostic test for malaria (WHO-AFRO andCDC, 2001).Measurement ToolsIntegrated disease surveillance systems; routinehealth information systemsWhat It MeasuresMalaria is a major public health problem in thedeveloping world. Each year there are at least 300million acute cases of malaria globally, which resultin more than one million deaths. Most of thesedeaths occur in sub-Saharan Africa. Malaria isAfrica’s leading cause of under-five mortality(20%) and constitutes 10% of the continent’soverall disease burden. This indicator measuresachievement towards the Roll Back Malaria(RBM) goal of halving malaria mortality for theAfrican population by 2010 (WHO, 2000c).How to Measure ItThe data requirement is the number of malariacases, severe and uncomplicated (probable andconfirmed), among children aged 0-59 months ina defined population. Many countries report onlylaboratory-confirmed cases, but many in Sub-Saharan Africa report clinically diagnosed casesas well. Some countries occasionally conductmortality surveys, and most countries with a highproportion of malaria deaths carry out routinesurveillance in “endemic” areas. However, in manycases, surveillance systems function poorly.Diarrhea, Acute Respiratory Infection, and Fever295

Strengths and LimitationsA number of issues need to be mentioned. First,this indicator reflects the overall magnitude of theproblem of malaria cases but does not offer aprecise estimate because of serious underreportingfrom surveillance data. Second, because thisindicator is reported as a number rather than as aproportion, countries with lower rates of malariadeaths but larger populations will rank ahead ofcountries with proportionately higher malariadeaths rates. Third, aggregate figures at a nationallevel may disguise pockets of high risk in certainregions.The indicator is to be interpreted with caution asthe level of completeness and timeliness ofreporting can vary across place and time.Surveillance systems reporting the number ofmalaria cases should also give the percentcompleteness of reporting. When examiningtrends in national malaria cases, it must beestablished whether a decrease in the number ofmalaria cases is perhaps due to incompletereporting or a decrease in the true cases resultingfrom education programs and malariainterventions.296Chapter 7

-FMALALARIAARIA DEAEATHRATEAMONG UNDERNDER-F-FIVESDefinitionThe number of deaths due to malaria (probableand confirmed) among children aged 0-59 monthsin a specified period per 100,000.Numerator: The number of deaths due tomalaria (probable and confirmed) amongchildren aged 0-59 months in a specifiedperiod x 100,000.Denominator: The total number of childrenaged 0-59 months in the same period.WHO standard case definitions are as follows(WHO and UNICEF, 2005):“Probable malaria death:” death of a person whowas diagnosed with probable severe malaria.“Probable severe malaria:” a person who requireshospitalization for signs and/or symptoms ofsevere malaria and receives antimalarial treatment.“Confirmed malaria death:” death of a person whowas diagnosed with severe malaria, with laboratoryconfirmation of diagnosis.“Confirmed severe malaria:” a person requiringhospitalization for signs and/or symptoms ofsevere malaria who receives antimalarial treatmentwith laboratory confirmation of diagnosis.“Severe malaria:” the signs and symptoms of severemalaria are the following: cerebral malaria;convulsions; circulatory collapse; abnormalbreathing (pulmonary edema or respiratorydistress syndrome); jaundice; macroscopichemoglobinuria; extreme weakness/prostration(inability to sit without support, or inability to feed,depending on the age of the child); renalimpairment; severe anemia; and hypoglycemia.These manifestations can occur singly, or morecommonly, in combination in the same person.The most common and important manifestationsof severe malaria in children are cerebral malaria,severe anemia, respiratory distress, andhypoglycemia (WHO, 2000b; WHO/AFRO,2001).Measurement ToolsAdministrative data; community/demographicsurveillance; vital registration; verbal autopsy;household surveysWhat It MeasuresThis indicator measures deaths from malariaamong children under the age of five years.Malaria accounts for one in five of all childhooddeaths in Africa. Measurement of the malariadeath rate among children is important for severalreasons. These include: (1) to establish the relativepublic health importance of malaria as a cause ofunder-five mortality; (2) to evaluate healthinterventions aimed at reducing under-fivemortality from malaria when these interventionsare being introduced in a limited geographic areaon a trial basis; (3) to investigate the circumstancessurrounding the deaths of children in order toidentify ways to reduce unnecessary death; and (5)to facilitate research into factors associated withunder-five mortality from malaria.How to Measure ItData come from administrative sources,community/demographic/integrated diseasesurveillance, household surveys, and vital statisticsregistrations. Administrative data are derived byhealth ministries from routine health informationsystems. Vital statistics registration systems collectdata on cause of death, including deaths causedby malaria. Good quality information requires thatDiarrhea, Acute Respiratory Infection, and Fever297

death registration be near universal, that the causeof death be reported routinely on the death record,and that it be determined by a qualified observeraccording to the International Classification ofDiseases (ICD). Such information is not generallyavailable in developing countries.This indicator can be measured at the communityor health facility, district, and national levels. Inareas where medical certification of the cause ofdeath is rare, verbal autopsy can be used to identifythe causes of death among infants and children.In verbal autopsy, two different methods can beused to get a verbal account of the cause of death:an open-ended history of final illness and closeendedquestions. In the open-ended history, thecaregiver or next-of-kin is asked to tell about theevents leading up to the child’s death in their ownwords and probed to follow-up on particularaspects. Then the descriptive account is reviewedby medical experts who then code the interviewin terms of cause of death. Close-ended questionsask whether specific symptoms and signs werepresent during the final illness (WHO, 1999b).Some DHS and MICS also provide data on causesof under-five mortality. Demographic surveillance,where all deaths are reported on a regular basisthroughout the year (often once every two weeks),can also be used for identifying deaths. WHOalso produces model-based estimates of malariaspecificmortality.The overall goal set by the Abuja Declaration isto halve the malaria mortality for Africa’s peopleby 2010 (WHO, 2000a).Strengths and LimitationsObtaining reliable data on malaria mortality isdifficult, partly because it is difficult to assigncauses of death. Misclassification of the cause ofdeath not only affects estimates of levels of malariamortality over time, but it also affects differencesin malaria mortality rates between two populationgroups, and difference in mortality due to all othercauses. In mortality surveys, the accuracy of theindicator depends on the ability of respondents todescribe the final illness and on the way in whichdiseases are understood and described in thecommunity.At the community level, data collection is oftenretrospective and relies on verbal autopsy, ratherthan on a clinically determined cause of death.These factors contribute to recall bias and makethe validity and reliability of the data questionable.In some cultures, death is a taboo subject. Thismakes asking questions about deaths problematic,which could lead to an underestimation ofmortality. In addition, very large sample sizes arerequired to calculate under-five deaths rates. Inmost sites, therefore, it will be impossible togenerate a malaria-specific mortality rate.Vital registration systems often do not havesufficient coverage to provide accurate data ofcause-specific mortality in developing countries.As demographic surveillance tends to cover limitedgeographic areas, underlying cause-specificmortality in these areas cannot necessarily begeneralized to wider populations.One limitation of the indicator is that the deathof a child is commonly the result of more thanone cause. Some verbal autopsy questionnaires,such as that developed by WHO, Johns HopkinsSchool of Hygiene and Public Health, and theLondon School of Hygiene and TropicalMedicine, allow for multiple causes of death, whileothers only allow one cause of death. Wheninterpreting this indicator, it is important to knowwhether multiple causes of death are allowed forin the coding since the expected proportions ofdeath for each cause, including malaria, willgenerally be higher when multiple causes of deathare allowed.298Chapter 7

RefererencesencesAlonso PL, Lindsay SW, Armstrong-Schellengberg JRM, Keita K, Gomez P, Shenton FC, Hill AG, DavidPH, Fegan G, Cham K, and Greenwood BM. (1993). A Malaria Control Trial Using Insecticide-TreatedBednets and Targeted Chemoprophylaxis in a Rural Area of Gambia, West Africa. The Impact of theInterventions on Mortality and Morbidity from Malaria. Transactions of the Royal Society of Tropical Medicineand Hygiene, 87(supp. 2):37-44.Bateman O, Jahan R, Brahman S, Zeitlyn S, and Laston S. (1995). Prevention of Diarrhea through ImprovingHygiene Behaviors: The Sanitation and Family Education (SAFE) Pilot Project Experience. Bangladesh: CARE-ICDDR,B.Billig P, Bendahmane D, and Swindale A. (1999). Water and Sanitation Indicators Measurement Guide.Washington, DC: Food and Nutrition Technical Assistance (FANTA) Project.Black RE. (1993). Persistent Diarrhea in Children in Developing Countries. Pediatric Infectious DiseaseJournal, 12:751-761.Black RE, Morris SS, and Bryce J. (2003). Where and Why Are 10 Million Children Dying Every Year?Lancet, 361(9376):2226-34.Curtis V, and Cairncross S. (2003). Effect of Washing Hands with Soap on Diarrhoea Risk in the Community:a Systematic Review. Lancet Infect Dis, 3(5):275-81.EHP. (2004). Hygiene Improvement: Guidelines for Household and Community Levels. Washington, DC: TheEnvironmental Health Project.Esrey SA, Potash JB, Roberts L, and Shiff C. (1991). Effects of Improved Water Supply and Sanitation onAscariasis, Diarrhoea, Dracunculiasis, Hookworm Infection, Schistosomiasis, and Trachoma. Geneva, Switzerland:World Health Organization. Bulletin of the World Health Organization, 69(5):609-21.Howard G, and Bartram J. (2003). Domestic Water Quantity, Service Level and Health. Geneva, Switzerland:World Health Organization. WHO/SDE/WSH/03.02. Retrieved from http://www.who.int/water_sanitation_health/diseases/wsh0302/en/Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS, and Bellagio Child Survival Study Group. (2003).How Many Child Deaths Can We Prevent This Year? Lancet, 362(9377):65-71.Kosek M, Bern C, and Guerrant RL. (2003). The Global Burden of Diarrhoeal Disease, as Estimated fromStudies Published Between 1992 and 2000. Geneva, Switzerland: World Health Organization. Bulletin ofthe World Health Organization, 81(3):197-204.Lengeler C, and Snow RW. (1996). From Efficacy to Effectiveness: Insecticide-Treated Bednets in Africa.Lancet, 74(3):325-332.Remme JH, Binka F, and Nabarro D. (2001). Toward a Framework and Indicators for Monitoring RollBack Malaria. American Journal of Tropical Medicine and Hygiene, 64(1-2 Suppl):76-84.Diarrhea, Acute Respiratory Infection, and Fever299

UNICEF. (1999). The Household and Community Component of IMCI: A Resource Manual on Strategies andImplementation Steps. Nairobi, Kenya: UNICEF ESARO.United Nations. (2000). Resolution 55/2: Millennium Declaration. New York: United Nations, GeneralAssembly. Retrieved from http://www.un.org/millennium/declaration/ares552e.htmVictora CG, Bryce J, Fontaine O, and Monasch R. (2000). Reducing Deaths from Diarrhoea Through OralRehydration Therapy. Geneva, Switzerland: World Health Organization. Bulletin of the World HealthOrganization, 78(10):1246-55.Whiteford L, Laspina C, and Torres M. (1996). Monitoring the Effect of Behavior Change Activities onCholera: A Review in Chimborazo and Cotopaxi, Ecauador. Arlington, VA: Environmental Health Project.Activity Report #25. Retrieved from http://www.dec.org/pdf_docs/PNABZ379.pdfWHO. (1999). IMCI Indicators, Monitoring and Evaluation. Geneva, Switzerland: World HealthOrganization. WHO/CHS/CAH/98.1K REV.1. Retrieved from http://www.who.int/child-adolescenthealth/New_Publications/IMCI/WHO_CHS_CAH_98.1/Rev99.K.pdfWHO. (1999a). The Evolution of Diarrhoeal and Acute Respiratory Disease Control at WHO: Achievements1980-1995 in Research, Development, and Implementation. Geneva, Switzerland: World Health Organization.WHO/CHS/CAH/99.12. Retrieved from http://www.who.int/child-adolescent-health/New_Publications/CHILD_HEALTH/WHO_CHS_CAH_99.12.pdfWHO. (1999b). A Standard Verbal Autopsy Method for Investigating Causes of Death in Infants and YoungChildren. Geneva, Switzerland: World Health Organization. WHO/CDS/CSR/ISR/99.4. Retrievedfrom http://www.who.int/csr/resources/publications/surveillance/WHO_CDS_CSR_ISR_99_4/en/WHO. (2000). Roll Back Malaria: Framework for Monitoring Progress and Evaluating Outcomes and Impact.Geneva, Switzerland: World Heath Organization. WHO/CDS/RBM.2000.25. Retrieved from http://rbm.who.int/cmc_upload/0/000/012/168/m_e_en.pdfWHO. (2000a). The Abuja Declaration and the Plan of Action: An Extract from the African Summit on RollBack Malaria, Abuja, 25 April 2000. Geneva: World Health Organization. WHO/CDS/RBM/2000.1.Retrieved from http://mosquito.who.int/docs/abuja_declaration.pdfWHO. (2000b). Management of Severe Malaria: A Practical Handbook, 2nd Ed. Geneva, Switzerland:World Health Organization. Retrieved from http://www.mosquito.who.int/docs/hbsm.pdfWHO. (2000c). Malaria in Africa. Roll Back Malaria Information Sheet No. 3. Retrieved March 2005, fromhttp://www.rbm.who.int/cmc_upload/0/000/015/370/RBMInfosheet_3.htmWHO/AFRO. (2001). Severe Malaria in the African Region: Results of a Multicentre Study. Belvedere,Zimbabwe: WHO/AFRO. Malaria: Liaison Bulletin of the Malaria Programme. WHO/AFRO. Vol. 4,No. 2. Retrieved from http://www.afro.who.int/malaria/bulletins/2001-04_vol4-2.pdf300Chapter 7

WHO, and UNICEF. (2000). Global Water Supply and Sanitation Assessment 2000 Report. Geneva,Switzerland: World Health Organization. Retrieved from http://www.who.int/docstore/water_sanitation_health/Globassessment/GlobalTOC.htmWHO, and UNICEF. (2004). WHO/UNICEF Joint Statement: Clinical Management of Acute Diarrhoea.Geneva, Switzerland: World Health Organization. WHO/FCH/CAH/04.7; UNICEF/PD/Diarrhoea/01.WHO, and UNICEF. (2005). World Malaria Report 2005. Geneva, Switzerland: World HealthOrganization. WHO/HTM/MAL/2005.1102.WHO-AFRO, and CDC. (2001). Technical Guidance for Integrated Disease Surveillance and Response in theAfrican Region. Retrieved from http://www.cdc.gov/epo/dih/Eng_IDSR_Manual_01.pdfZinc Investigators' Collaborative Group. (2000). Therapeutic Effects of Oral Zinc in Acute and PersistentDiarrhea in Children in Developing Countries: Pooled Analysis of Randomized Controlled Trials. AmericanJournal of Clinical Nutrition, 72:1516-22.Zinc Investigators' Collaborative Group. (1999). Prevention of Diarrhea and Pneumonia by ZincSupplementation in Children in Developing Countries: Pooled Analysis of Randomized Controlled Trials.Journal of Pediatrics, 135:689-97.Chapter 7. Diarrhea, Acute Respiratory Infection, and Fever301

8GROWTHMONIONITORORINGAND NUTRTRITIONIONIndicators:• Sick child checked for three danger signs (cross-referenced in Chapter six)• Sick child’s weight checked against a growth chart (cross-referenced inChapter six)• Sick child under two years of age assessed for feeding practices (crossreferencedin Chapter six)• Caretaker of sick child is advised to give extra fluids and continue feeding(cross-referenced in Chapter six)• Proportion of hospitals and maternity facilities designated as baby friendly(cross-referenced in Chapter three)• Exclusive breastfeeding rate (cross-referenced in Chapter three)• Timely initiation of breastfeeding• Complementary feeding rate• Mean dietary diversity of foods consumed by children aged 6-23 months• Proportion of children aged 6-23 months with good young child feedingpractices• Proportion of households with access to essential handwashing supplies(cross-referenced in Chapter seven)• Proportion of households where the caretaker of the youngest child underfive reported appropriate handwashing behavior (cross-referenced inChapter seven)• Sick child aged 6-23 months is offered increased fluids and continued feeding• Proportion of children living in households using adequately iodized salt• Proportion of children aged 12-59 months who were dewormed in the pastsix months• Prevalence of night blindness in children• Vitamin A supplementation• Vitamin A deficiency (serum retinol concentration)• Proportion of children aged 6-59 months with anemia• Low weight-for-height/length (wasting)• Low height/length-for-age (stunting)• Low weight-for-age (under weight)• Under-five mortality rate (cross-referenced in Chapter nine)

HAPTER 8. GCHAPTER8. GROWTHMONITONITORINORING AND NUTRITIONCauses of Child Malnutrition and RelevantInterventionsThe widely accepted ‘framework’ for childmalnutrition was provided by UNICEF in1990 (UNICEF, 1990) and serves as a guide forassessing and analyzing the causes of childmalnutrition. The framework, presented in Figure8.1 on the following page, highlights the multisectoralnature of the causes of malnutrition, whichencompass inadequate access to food, inadequatecare for mothers and children, and insufficientheath services. The causes of malnutrition are alsocategorized as immediate, underlying, and basic.Basic causes of malnutrition include inadequateeducation, the political and ideologicalsuperstructure, the economic structure, and theresource base. These factors influence theunderlying causes – food, care, and health – whichaffect directly, in turn, the two primary proximatedeterminants of nutritional status and survival,namely dietary intake and disease.Inadequate dietary intake and infection constitutea vicious cycle that contributes to approximatelyhalf of the morbidity and mortality levels indeveloping countries. Malnutrition leads tocompromised immunity which results in greaterincidence, severity and duration of disease,impaired motor and cognitive development, andreduced physical fitness. For girls, poor nutritionhas an intergenerational component. Childrenborn to malnourished mothers are more likely tobe low birth weight and to die as infants. If thesechildren are female and survive, they are likely tobe stunted in adulthood and to give birth to lowbirth weight babies if something is not done tobreak the cycle. The framework does notemphasize the distributional aspects ofmalnutrition – of resources in society, for instance– nor the dynamic interaction between the factors,notably between malnutrition and infection. ThisChapter 8. Growth Monitoring and Nutrition“malnutrition-infection cycle” (Tomkins &Watson, 1989) accounts for much of child illhealth and mortality in developing countries(Pelletier, 1994; WHO, 1996).The framework helps programs to identify themost appropriate mixture of child nutritioninterventions. In general, child nutrition programsaim to affect immediate and underlying causes ofmalnutrition and cut into the vicious cycle ofmalnutrition-infection, often through acombination of community organizations andservice delivery (Tontisirin & Winichagoon, 1999)plus specific micronutrient interventions (Mason,Mannar, & Mock, 1999). The content ofcommunity-based program activities is fairly wellestablished and includes antenatal care,breastfeeding support, complementary feeding,growth monitoring, micronutrientsupplementation, supplementary feeding,immunization, de-worming, and health referral.Additional activities not always identified withnutrition (but which affect it) include day-care,water/sanitation, and microcredit (see Mason etal, 2001b, pp. 53-54). The precise scope ofprograms aimed at child health and nutritiondepends, of course, on local circumstances.However, the main focal points are food, health,and care.Improving “household food security” should bean essential component of interventions that seekto promote adequate access to food. Indeed, mostpolicies and programs aimed at improving foodaccess are the same as those aimed at reducingpoverty – which is not surprising since about twothirdsof the income of poor households goes tofood. However, many activities undertaken toimprove household security may be outside thescope of health/nutrition programs. Interventionsmay be introduced from several sectors, forexample, agriculture, education, and livelihoods.303

Figure 8.1. Conceptual framework for the causes of malnutrition in society.From a monitoring and evaluation standpoint,nutrition programs that seek to improve householdfood security may have to draw on indicators fromother sectors. Furthermore, if programs areintroduced as packages of services, it may bedifficult to establish that observed changes innutrition outcomes are attributable to a particularintervention. These issues will be discussed ingreater detail in subsequent sections. For thepresent purposes, the most common componentsof child nutrition programs are used as a guide.They include both interventions aimed at generalmalnutrition (or caloric deficit measured bygrowth) and those aimed at specific micronutrientdeficiencies.Conceptual FrameworkFigure 8.2 on the following page presents aconceptual framework for monitoring andevaluating child nutrition programs. The causalpathways by which nutrition interventionsimprove child health and nutritional status areoutlined and outputs and outcomes for commoncomponents of nutrition programs specified.Activities conducted in various functional areasof nutrition programs contribute to strengthenedservices to support breastfeeding and childnutrition. For example, service providers must betrained to manage child-feeding problems whenthey arise. IEC campaigns could be undertaken304 Chapter 8

Figure 8.2. Conceptual framework for monitoring and evaluating child nutrition programs.305

to increase levels of knowledge regarding infantfeeding and child nutrition. As in the previouschapters, outputs are categorized into functionaloutputs and services output. Functional outputsmay include the availability of vitamin A capsules;the presence of health workers whose trainingincludes key nutrition elements; the availabilityof child feeding counseling and IEC materials;supervisory visits that include observation ofnutrition counseling; and so forth. Althoughmeasures relevant to the functional areas of childhealth programs are not presented here, they areimportant for tracking whether activities are beingimplemented as planned.The service outputs suggested by the monitoringand evaluation (M&E) framework recapitulateand expand on indicators developed by WHO andUNICEF for infant feeding and child nutrition.Measures of the quality of services encompass thenutrition content of sick childcare (routinescreening of sick children for visible signs ofwasting, edema, very low weight and illness;assessment of breastfeeding and complementaryfeeding practices; weighing of sick children andplotting their weight on a growth chart); regularweighing of well children; counseling of mothersabout infant and young child feeding; checkingof vitamin A supplementation whenimmunizations are given; integrating vitamin Asupplementation and guidance into communitybasedweighing session, and so forth.Commitment to the Baby Friendly HospitalInitiative is another aspect of the quality of carethat demonstrates institutional recognition of theimportance of breastfeeding and appropriatebirthing practices to support its initiation. Manyof the indicators of the quality of services areparticularly significant for evaluation purposesbecause they measure the implementation ofinterventions commonly considered key tosuccessful child feeding promotion and support.In the long term, the results of child nutritionactivities are manifested in the reduced prevalenceof micronutrient deficiency; less growth faltering;the reduced incidence, severity, and duration ofillness; and lower infant and child mortality. Childmalnutrition is often measured by growth, to thepoint that low anthropometric measurement hasbecome almost synonymous with malnutrition.This phenomenon used to be referred to as caloricor energy deficit. However, since growth failureis non-specific to cause, and because specificnutrient deficiencies can cause growth failure,general malnutrition (from unspecified causes) isthe term that is increasingly being used to refer togrowth failure (Beaton et al., 1990). Otheroutcomes that may stem indirectly from nutritionactivities but are not included in the frameworkinclude cognitive and motor development, physicalfitness, and overall functioning.A crucial aspect of this M&E framework is therecognition that the effects of program activitiesdepend on contextual factors. Contextual factorsinclude women's status, literacy, andenvironmental and political factors that cannoteasily be changed by a project's activities, but mayfundamentally modify their effects (see Mason etal., 2001b, pp 42-49). In M&E terms, this meansthat contextual factors need to be measured, andstatistical methods used to control for them, toarrive at a better understanding of how differentgroups are influenced by program actions.Although many contextual factors arestraightforward to measure (e.g., female education,proxies for socioeconomic status [SES] such ashousing, access to services, etc.) and are includedin surveys such as the DHS and MICS, thesefactors are often ignored in program planning.Coverage, Intensity, Targeting, and ContentThe effect of an intervention on child nutritionoutcomes can be seen as depending on four factors:coverage, intensity, targeting, and content.Coverage refers to the proportion either of thepopulation, or the needy in the population (e.g.underweight children) that participate in theprogram. Coverage can be estimated fromprogram-derived data alone, if the size of the targetpopulation is known from other sources.306 Chapter 8

Targeting means focusing an intervention onspecific groups based on their age, socioeconomiccharacteristics, and/or geographic location.Targeting may address different dimensions of theexpected results and be expressed in terms ofquantity (how much), quality (how good), orefficiency (least cost) values to be achieved withina specific time frame. When a program's progressis to be measured in terms of its effects on people,the expectation is that changes in nutritionoutcomes, health, and behavior would be morebeneficial among program participants than in thegeneral population; hence the ratio of theprevalence of desired nutrition outcomes amongthe intended beneficiaries to the prevalence in theoverall population should be greater than 1. Thiswould signify that the program is preferentiallygetting to the malnourished. In practice, whenthis indicator is calculated by administrative area,it usually emerges that programs are only slightlytargeted, if at all, towards the worst-off.The idea underlying intensity is that it is necessaryto reach a certain level of exposure to anintervention for there to be an impact. Intensitycan be measured in terms of expenditure perparticipant per year or resources per capitarequired, which can be estimated in concrete termsas numbers of people to be trained, equipped, orotherwise supported. For community-basednutrition programs, key resource requirements mayinclude increased numbers of village-level workersor volunteers (mobilizers), supported bysupervisory staff (facilitators). Experiences inThailand have shown that the achievement ofimpact in community-based nutrition programsrequires ratios of mobilizers to populations ofapproximately 1:100, ratios of mobilizers tohouseholds of approximately 1:10-20, and ratiosof facilitators to mobilizers of approximately 1:10-20.Intensity estimates should be related to rates ofchange in the level of malnutrition, usuallymeasured by underweight prevalence in childrenaged 0-5 years. The underlying rate of change ofunderweight prevalence is around 0.5 percentageGrowth Monitoring and Nutritionpoints per year (pp/yr) – more in SE Asia (up to 1pp/yr) and less in Sub-Saharan Africa (aboutstatic). For community-based programs, with theusual content of activities and adequate resources(i.e. intensity), an additional 1 to 1.5 pp/year ofimprovement in underweight prevalence can beexpected (Swindale, Deitchler, Cogill, andMarchione, 2004). Though experience issomewhat limited, it has been shown that anintensity of $5-15/child/year should lead to anaccelerated decrease in the child underweightprevalence of approximately 1-1.5 percentagepoints/year. This level of intensity may be beyondthe range of some projects. However, if programinput is low, it is questionable as to whetherprograms can actually have an impact. Thus,investing insufficient resources in nutritioninterventions may be wasteful (see Mason &Habicht, 1984). One option would be to focusavailable resources by targeting until a level isreached where impact can be achieved.Given adequate coverage and intensity, theremaining factor for success is the content of theprogram – that is, the activities supported and theextent to which they are matched to local causesof malnutrition and existing opportunities andorganizational conditions. An incorrectly assumedrelation of a problem to its causes may lead toselecting interventions that cannot be effective.For example, low dietary iron might be taken asthe cause of anemia when the real issue is wormsor malaria. Thus, program content needs to beassessed for its relevance and likely effectiveness.Efficacy and EffectivenessIn general, nutrition interventions suffer from alack of rigorous evaluation. Small-scale studiesshowing the efficacy of different interventions onnutritional status are perhaps sufficient to establishwhat factors are needed for success (e.g., Pinstrup-Andersen et al., 1993). The term “efficacy” is oftenused to refer to the effect of an intervention onthe outcome at the pilot level, where conditionsare carefully controlled (and the change inoutcome can be more rigorously ascribed to the307

intervention and referred to as impact). Manygovernments are now interested in applyinginterventions on a large scale. For large-scaleprograms (which should be based on the efficacyexperience) the change in outcome ascribed to theprogram (i.e. impact) is referred to as“effectiveness.”The distinction between these concepts isimportant. It cannot be assumed that anefficacious pilot project can always be scaled up todirectly lead to an effective large-scale program.There are many well-known reasons for this. Thepilot community or area may represent idealconditions that are far from the reality of otherregions, communities, or areas. Therefore, the pilotprogram may not work as well if implemented ina different region or community. Anothercommon explanation is that strong leadership inpilot projects is an important factor contributingto program success. However, individuals orgroups who represent leadership, interest,enthusiasm, determination, and assertivenesscannot easily be reproduced or identified on alarge-scale.How does one choose an appropriate evaluationdesign to measure program efficacy oreffectiveness? The complexity of the evaluationdesign depends on who the decision maker is andon what types of decisions will be taken as aconsequence of the evaluation findings. Differentdecision makers demand not only different typesof information but also vary in their requirementsof how informative and precise the findings mustbe. Habicht and his colleagues (1999) haveproposed a framework for choosing an appropriateevaluation design based on two considerations.The first consideration concerns the indicators:What does the program want to measure: serviceprovision or utilization, program coverage, orimpact? The second consideration refers to thetype of inference: How confident does theprogram or decision maker want to be that anyobserved effects were in fact due to the programactivities/intervention(s)?Based on these considerations, evaluations may becategorized into three groups: adequacy,plausibility, and probability evaluations (Habichtet al., 1999). An adequacy evaluation answers thequestions: Did the expected changes occur? Aplausibility evaluation wants to find out whetherthe program seems to have an effect above andbeyond other external influences. Finally, aprobability evaluation aims at ensuring that thereis only a small known probability that thedifference between program and control areas aredue to confounding problems, bias, or chance.Probability evaluations, also known as impactevaluations, require randomization of treatmentand control activities and are the gold standard ofefficacy research.When considering M&E needs and design, it isimportant that an individual program finds outthe following:• Has the effectiveness of similar large-scaleprograms been established (plus, in whatdirection and how large was the effect)? If alarge-scale program is based on anintervention with proven efficacy in field trials,there would be little need for a probabilityevaluation (involving randomization oftreatment and control).• If the outcomes of similar large-scale programsare positive, then it may be enough to knowthat the program is being implemented asintended and the trend in outcome is in theright direction.• But if not, then it may be important to have acarefully designed evaluation that can elicitplausible evidence of actual impact.Methodological Challenges of EvaluatingInfant and Young Child FeedingThe major challenge of evaluating child nutritionprograms is establishing a causal relationshipbetween the intervention in question and thedesired outcome. This occurs because of thecomplexity of the interventions necessary to308 Chapter 8

improve children's nutritional status. The fact thatinterventions are rarely carried out in isolation ofother health programs complicates establishingdefinitive cause and effect relationships. In thissection, we focus on the specific challenges ofmonitoring and evaluating programs aimed atimproving infant and young child feeding anddraw on Bertrand and Escudero's (2003) summaryfor such interventions. Readers are referred toWHO (2001a; 2001b) for a review of themethodological challenges of monitoringprograms aimed at the elimination of iodinedeficiency disorders and the prevention andcontrol of iron-deficiency anemia.Infant feeding behavior data relies upon accurate agedata of the infantAlthough evaluators may track healthinterventions with only a general reference to thechild's age (e.g., less than one year), trackingbreastfeeding practices requires an accurateassessment of the infant's age. Interviewers canascertain the age by first asking the mother forthe infant's birth date and then by confirming thebirth date with a child health card or other officialregistry of the child's birth date.24-hour recall data tend to overestimate thepercentage of infants who have been exclusivelybreastfed since birthA 24-hour recall measure reflects currentbreastfeeding status and may cause the proportionof exclusively breastfed infants to be slightlyoverestimated, since some infants who consumeother liquids irregularly may not have receivedthem in the 24 hours before the survey. WHO'sIndicators for Assessing Breastfeeding Practices(WHO, 1991), Wellstart International's Tool Kitfor Monitoring and Evaluating BreastfeedingPractices and Programs, and DHS reports allcalculate the exclusive breastfeeding rate (EBR)using the 24-hour recall method. Using crosssectionalsurveys, one can obtain the best estimatesof exclusive breastfeeding from current status datathat include all births within a specified timeperiod. The advantage of this approach is that itis not subject to recall error. Evaluators shouldthen interpret the measure as the percentage ofinfants who “are currently being exclusivelybreastfed” rather than the percentage that havebeen exclusively breastfed since birth.Large sample sizes are needed to detect change inbreastfeeding practices, but infants represent a smallproportion of the population.Any assessment of behavioral change in infantfeeding requires attention to sample size. Thesample size depends on both the magnitude ofchange and on the prevalence of the condition orpractice. The detection of relatively small changes(e.g., five to ten percentage points) over time inbreastfeeding or other infant feeding behaviorsrequires large sample sizes. By contrast, simplemonitoring of infant feeding practices does notrequire a specific sample size and can be very usefulin tracking ongoing project outreach. However,monitoring neither allows for a rigorous evaluationof change, nor measures actual prevalence of thisbehavior because of small unrepresentativesamples.Breastfeeding questions typically require more thana “yes” or “no” response.Multiple factors define whether breastfeeding isoptimal, including the exact liquids and foods, ifany, given in the preceding 24 hours and the ageof the infant. Ideally, this list of liquids and foodswill be comparable to those included in the coreDHS questionnaire, with additional items thatreflect local food preferences and food availability.The data needed to calculate exclusivebreastfeeding and timely complementary feedingrates require that the interviewer ask therespondent a series of questions about all foodsand liquids given within the previous 24-hours.This line of questioning requires more than a “yes”or “no” response, thus increasing the likelihood ofinterviewer or respondent error. Interviewersshould undergo intensive training on this set ofitems.Growth Monitoring and Nutrition309

The accepted standard complementary feedingindicator reflects general dietary intake of solid andsemi-solid foods during a specified period only andfails to capture several important elements of optimalcomplementary feeding.Complementary feeding, a highly complex issue,involves factors such as the quantity and qualityof food, frequency and timeliness of feeding, foodhygiene, and feeding during/after illness. Programpersonnel at the country level must consider thesefactors when they try to address the problems ofinfant and young child feeding in the local context.The standard complementary feeding (CF)indicator fails to account for program-specific orcontext-specific feeding recommendationsregarding the frequency, quality, or quantity offoods given during the preceding 24 hours.The lack of consensus on recommended youngchild feeding practices has hindered progress indeveloping appropriate child feeding indicators.Recent efforts by WHO to develop anddisseminate consistent and internationallyaccepted guidelines and comprehensive reviews ofscientific knowledge regarding complementaryfeeding (Dewey and Brown, 2003; WHO/UNICEF, 1998) have led to the 10 “GuidingPrinciples” for complementary feeding of thebreastfed child (PAHO/WHO, 2003). Theseguiding principles are presented in Table 8.1 onpage 312 and form the basis for current efforts togenerate indicators of appropriate feeding ofchildren aged six through 23 months (Arimondand Ruel, 2003).Selection of IndicatorsThe indicators presented in this chapter are fewin number, fairly easy to measure and interpret,and operationally useful. Moreover, they havebeen field-tested, are consistent with worldwidebreastfeeding and complementary feeding goals,and can be obtained from available DHS, MICS,or KPC data. The indicators include measures ofthe quality of the nutrition component of sick childcare, water, sanitation, and hygiene, infant andyoung child feeding, micronutrient deficiency,vitamin A supplementation, and anthropometry,and can be grouped into the three categoriespresented below:Service output• Sick child checked for three danger signs• Sick child’s weight checked against a growthchart• Sick child under two years of age assessed forfeeding practices• Caretaker of sick child is advised to give extrafluids and continue feeding• Proportion of hospitals and maternity facilitiesdesignated as baby friendlyOutcome• Exclusive breastfeeding rate• Timely initiation of breastfeeding• Complementary feeding rate• Mean dietary diversity of children aged 6-23months• Proportion of children aged 6-23 months withgood young child feeding practices• Proportion of households with access toessential handwashing supplies• Proportion of households where the caretakerof the youngest child under five reportedappropriate handwashing behavior• Sick child aged 6-23 months is offeredincreased fluids and continued feeding• Proportion of children living in householdsusing adequately iodized salt• Proportion of children aged 12-59 monthswho were dewormed in the past six months• Prevalence of night blindness in children• Vitamin A supplementation• Low weight-for-height/length (wasting)310 Chapter 8

Impact• Vitamin A deficiency (serum retinolconcentration)• Proportion of children aged 6-59 months withanemia• Low height/length-for-age (stunting)• Low weight-for-age (under weight)• Under-five mortality rateThe indicators pertaining to infant and youngchild feeding during the first two years of life canbe further broken down into two groups: (1) thoseconcerning breastfeeding behaviors during the firstsix months of life, and (2) those concerning theintroduction of complementary foods whilemaintaining breastfeeding beginning at sixmonths. Together, these age groups represent thecontinuum of infant and child nutrition care inthe first two years of life. The child feedingindicators for these two age groups reflect expertconsensus as to the optimal period for exclusivebreastfeeding, as well as for the introduction ofcomplementary foods to an infant's diet. Inpopulation-based surveys, measuring these infantfeedingindicators requires the sampling of infants0-5 months of age and 6-23 months of age.The descriptions of some of the indicatorspresented in this chapter draw heavily from or areguided by previous work by Arimond and Ruel(2003), Bertrand and Escudero (2003), Cogill(2003), the World Health Organization, andUNICEF. Programs can use these indicators tomonitor and evaluate behavior changeinterventions in the context of an experimental orquasi-experimental design. Programs can alsocalculate these indicators from routine statisticsfor tracking behaviors among clients, and formeasuring the long-term outcomes of specificinterventions. In certain circumstances, additionalindicators may be warranted. These can beformulated at the national, regional, or district levelto reflect special needs and concerns.Growth Monitoring and Nutrition311

Table 8.1. PAHO/WHO guiding principles for complementary feeding of the breastfed child(1) Duration of exclusive breastfeeding and age of introduction of complementary foodsPractice exclusive breastfeeding from birth to six months of age, and introduce complementaryfoods at six months of age while continuing to breastfeed.(2) Maintenance of breastfeedingContinue frequent, on-demand breastfeeding until two years of age or beyond.(3) Responsive feedingPractice responsive feeding, applying the principles of psycho-social care. Specifically: (a) feedinfants directly and assist older children when they feed themselves, being sensitive to theirhunger and satiety cues; (b) feed slowly and patiently, and encourage children to eat, but do notforce them; (c) if children refuse many foods, experiment with different food combinations,tastes, textures, and methods of encouragement; (d) minimize distractions during meals if thechild loses interest easily; (f) remember that feeding times are periods of learning and love –talk to children during feeding, with eye-to-eye contact.(4) Safe preparation and storage of complementary foodsPractice good hygiene and proper food handling by: (a) washing caregivers’ and children’shands before food preparation and eating; (b) storing foods safely and serving foods immediatelyafter preparation; (c) using clean utensils to prepare and serve foods; (d) using clean cups andbowls when feeding children; and (e) avoiding the use of feeding bottles which are difficult tokeep clean.(5) Amount of complementary foodsStart at six months of age with small amounts of food and increase the quantity as the child getsolder, while maintaining frequent breastfeeding. The energy needs from complementary foodsfor infants with average breastmilk intake in developing countries are approximately 200 kcalper day at 6-8 months, 300 kcal per day at 9-11 months of age, and 550 kcal per day at 12-23months of age. In industrialized countries these estimates differ somewhat (130, 310, and 580kcal/d at 6-8, 9-11, and 12-23 months, respectively), because of differences in average breastmilkintake.(6) Food consistencyGradually increase food consistency and variety as the infant gets older, adapting to the infant’srequirements and abilities. Infants can eat pureed, mashed, and semi-solid foods beginning atsix months. By eight months, most infants can also eat “finger foods” (snacks that can be eatenby children alone). By 12 months, most children can eat the same types of foods as consumedby the rest of the family (keeping in mind the need for nutrient-dense foods, as explained in #8on the following page). Avoid foods that may cause choking (i.e., items that have a shape and/or consistency that may cause them to become lodged in the trachea, such as nuts, grapes, andraw carrots).312Chapter 8

Table 8.1. PAHO/WHO guiding principles for complementary feeding of the breastfed child (continued)(7) Meal frequency and energy densityIncrease the number of times that the child is fed complementary foods as s/he gets older.The appropriate number of feedings depends on the energy density of the local foods and theusual amounts consumed at each feeding. For the average healthy breastfed infant, meals ofcomplementary foods should be provided 2-3 times per day at 6-8 months of age and 3-4times per day at 9-11 and 12-24 months of age, with additional nutritious snacks (such as apiece of fruit or bread or chapatti with nut paste) offered 1-2 times per day, as desired. Snacksare defined as foods eaten between meals – usually self-fed, convenient, and easy to prepare. Ifenergy density or the amount of food per meal is low, or the child is no longer breastfed, morefrequent meals may be required.(8) Nutrient content of complementary foodsFeed a variety of foods to ensure that nutrient needs are met. Meat, poultry, fish, or eggsshould be eaten daily, or as often as possible. Vegetarian diets cannot meet nutrient needs atthis age unless nutrient supplements or fortified products are used (see #9 below). Vitamin A-rich fruits and vegetables should be eaten daily. Provide diets with adequate fat content.Avoid giving drinks with low nutrient value, such as tea, coffee, and sugary drinks such as soda.Limit the amount of juice offered so as to avoid displacing more nutrient-rich foods.(9) Use of vitamin-mineral supplements or fortified products for infants and motherUse fortified complementary foods or vitamin-mineral supplements for the infant, as needed.In some populations, breastfeeding mothers may also need vitamin-mineral supplements orfortified products, both for their own health and to ensure normal concentrations of certainnutrients (particularly vitamins) in their breastmilk. (Such products may also be beneficial forpre-pregnant and pregnant women.)(10) Feeding during and after illnessIncrease fluid intake during illness, including more frequent breastfeeding, and encourage thechild to eat soft, varied, appetizing, favorite foods. After illness, give food more often thanusual and encourage the child to eat more.Excerpted from PAHO/WHO (2003).Growth Monitoring and Nutrition313

SICICK CHILDCHECHECKEDFORTHREEDANANGERSIGNSPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children checked for the threegeneral danger signs.Numerator: Number of sick children aged 2-59 months seen who are checked for threedanger signs (is the child able to drink orbreastfeed; does the child vomit everything;has the child had convulsions).Denominator: Number of sick children aged2-59 months seen.Measurement ToolsService Provision Assessment (SPA); HealthFacility Assessment (HFA); supervision checklistWhat It MeasuresThis indicator measures the performance of oneof the tasks associated with the routine assessmentof sick children aged two months to five years:checking for general danger signs. The indicatorpresupposes that the health worker has beentrained in IMCI. Given this assumption, theindicator measures both the adequacy of IMCItraining to impart these skills and the ability ofthe trainees to assimilate and retain theinformation and skills over time.child has had convulsions during the presentillness. It is recommended that health workersnot rely completely on the caretaker’s report ofwhether the child is able to drink or breastfeedbut observe the mother while she tries tobreastfeed or to give the child something to drink.Strengths and LimitationsMeasuring this indicator is straightforward andcan be done in a routine basis during supervisoryvisits. Immediate feedback can be given to thehealth worker to improve future practice. Theindicator can help identify if health workers of aparticular health facility need refresher training.If no improvement in this indicator is seen overtime, program managers may wish to monitorhealth workers’ assessment of the presence of eachdanger sign separately to identify whether aparticular danger sign is not routinely checked.Sample QuestionsSample instructions from a checklist completedby an observer during a SPA are the following:o Record whether a provider asked about orwhether the caretaker mentioned any of thefollowing:(1) Whether the child is unable to drink orbreastfeed at all;(2) Whether the child vomits everything; and(3) Whether the child has had convulsionswith this sicknessHow to Measure ItMeasurement of this indicator requires directobservation of sick child consultations todetermine whether the health worker asked aboutor the caretaker reported each of the followingdanger signs: the child is unable to drink orbreastfeed, the child vomits everything, and theGrowth Monitoring and Nutrition315

SICICK CHILDHILD’S WEIGHTCHECHECKEDAGAINSAINST A GROWTHCHARHARTPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children aged 2-59 months whoare weighed the day they are seen and whose weightsare checked against a recommended growth chart.Numerator: Number of sick children aged 2-59months who are weighed the day they are seenand whose weights are checked against arecommended growth chart.Denominator: Total number of sick children aged2-59 months seen.Measurement ToolsSPA; HFA; supervision checklistWhat It MeasuresThis indicator measures health worker compliancewith IMCI guidelines for the routine assessment ofthe nutritional status of sick children by weighingthe child and plotting the weight on a recommendedgrowth chart. There are two main reasons for thisassessment. The first is to identify children withsevere malnutrition who are at increased risk ofmortality and need urgent referral. Second, anassessment of the nutritional status of sick childrenhelps to identify children with low weight-for-agewho may benefit from nutritional counseling. Allsick children should be assessed for malnutrition.How to Measure ItThe data are gathered by direct observation of sickchild consultation. The observer records whetherthe health worker weighed the child and plotted thechild’s weight on a recommended growth chart(usually a standard WHO or national growth chart).Strengths and LimitationsData for this indicator are easy to collect during sickchild observation or routine supervisory visits. Theygive a good indication of health worker compliancewith IMCI guidelines regarding the nutritionalassessment of all sick children. The indicator can beapplied at a specific interval post-training to thosewho attended IMCI training to evaluate theretention of this particular component of clinicalassessment skills. This may help identify healthworkers who need refresher training or health centersin which weighing of sick children and recordingthe weight on a growth chart are not enforced.Limitations to the use of observation for measuringquality of sick child assessment have already beendiscussed. These include “observation bias,” in thata health worker may abide to the guidelines morestrictly when he or she is conscious of beingmonitored. Another limitation of the indicatorpertains to variability between observers inmeasurement. This is hard to measure but can beassessed by having two independent observers rate asick child consultation and then comparing thedegree of agreement or disagreement in their ratings.Note that the indicator does NOT reflect the correctmeasurement of the child’s weight, whether thechild’s weight was accurately plotted on a growthchart, how effectively health workers interpret theinformation on the growth chart, or whether thehealth worker took an appropriate course of actionbased on insights from the growth chart. It is alsodifficult to tell from the indicator whether healthworkers are weighing/not weighing the children atall, or whether children are weighed but their weightsare not plotted on a growth chart.Growth Monitoring and Nutrition317

SICICK CHILDUNDERTWO YEARSOF AGEASSESSEDFORFEEDINEEDING PRARACTICESPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children under two years of agewhose caretakers are asked about breastfeeding,complementary foods, and feeding practicesduring this episode of illness.Numerator: Number of sick children undertwo years of age whose caretakers are asked ifthey breastfeed this child, whether the childtakes any other food or fluids other than breastmilk, and if during this illness the child’sfeeding has changed.Denominator: Number of sick children undertwo years of age seen.Measurement ToolsSPA; HFA; supervision checklistWhat It MeasuresThis indicator is a composite of the steps a healthworker must undertake in order to correctly assessa sick child’s feeding and counsel the mother tosolve any feeding problems that exist. Thus, itmeasures the correct assessment of a sick child’sfeeding. The IMCI guidelines require that all sickchildren under two years have a feeding assessmenteven if they have a normal Z-score (WHO, 1998).The standard deviation unit or Z-score is thesimplest way of making comparisons to thereference population. The Z-score is defined asthe difference between the value for an individualand the median value of the reference populationin the same age or weight, divided by the standarddeviation of the reference population. The medianis the value at exactly the mid-point between thelargest and smallest.How to Measure ItThe measurement of this indicator is based onobservations of the performance of health workerswith regard to the assessment of feeding practicesfor sick children who are brought to a healthfacility. The observer records on a questionnaireor checklist whether the health worker asked aboutor the care taker reported on whether the child isbreastfed, whether the child is taking any otherfoods or fluids, and whether feeding practices hadchanged during the illness. Routine supervisoryvisits can also yield data for measuring thisindicator.Exit interviews are a second option for collectingdata to measure this indicator. The caretaker isasked whether the health worker had asked him/her questions about breastfeeding, complementaryfeeding, and changes in the child’s feeding duringthe current illness.Only a sick child less than two years old who hasa “yes” answer on all three aspects of the assessmentof feeding practices is included in the numerator.The denominator is the total number of sickchildren under the age of two years who were seenby the health worker.Strengths and LimitationsMethodologically, this indicator is relatively easyto construct. The indicator could be aggregatedacross all health workers in a facility to calculatean average child feeding assessment score for agiven health facility. This is useful for identifyingfacilities that are performing below standards withGrowth Monitoring and Nutrition319

egard to the assessment of feeding practices. Alow score on the indicator is a fairly soundindication of the need for refresher training ortargeted supervision.There are, however, difficulties in interpretingchanges in this indicator. Lack of change in theindicator could mean deterioration in one or moreaspects of child feeding assessment counteractedby improvements in other aspects, or could merelyindicate a general lack of performanceimprovement in this component of sick childassessment. For purposes of planning andmonitoring, it may be useful to monitor changesin the individual components of the indicator thatare of most interest to program managers. Thus,programs may want to calculate a separateindicator for each task associated with theassessment of child feeding in order to identifywhich tasks are not routinely performed.Sample QuestionsThe following are sample questions from the SPA(dated March 2004):o Record whether a provider asked about orperformed other assessments of the child’shealth by doing any of the following:(1) Offer the child something to drink or askthe mother to put the child to the breast(to find out whether the child can drink);(2) Ask about normal breastfeeding practiceswhen the child is not ill; and(3) Ask about feeding or breastfeedingpractices for the child during this illness.320 Chapter 38

CARETARETAKERAKER OF SICICK CHILDIS ADVISEDTO GIVEEXTRAFLUIDSAND CONTINUEFEEDINEEDINGPriorioritity y Indicator for IMCI at Health-Facilitacility y LevelelDefinitionProportion of sick children with validatedclassifications who do not need urgent referralwhose caretakers are advised to give extra fluidand continue feeding.Numerator: Number of sick children withvalidated classifications, who do not needurgent referral, whose caretakers are advisedto give extra fluid and continue feeding.Denominator: Number of sick children withvalidated classifications, who do not needurgent referral.Validated classification: All sick children need tobe classified correctly by the health worker. Thevalidation of the health worker’s classificationneeds to done at the health center by an authorizedperson (basically a doctor or clinical supervisor)who reexamines a sample of cases seen by thehealth worker to see if a correct classification wasmade.Measurement ToolsSPA; supervisory checklistWhat It MeasuresThis indicator measures the extent to which healthworkers are complying with standards forcounseling a mother or caretaker of a sick childon the need to continue feeding and increase thechild’s fluid intake at home. Sick children needto increase their fluid intake during and afterillness to avoid dehydration. This is especiallyimportant for diarrheal illness but is also true forother illnesses that may make children less likelyto drink. Continued feeding (includingbreastfeeding) during illness shortens the durationof the illness episode and reduces the risks ofdehydration and growth faltering.How to Measure ItThe information required for this indicator canbe obtained during a health facility survey throughdirect observations of sick child consultations orexit interviews of caretakers. A health worker isincluded in the numerator if he or she scorespositively on both advice about the need tocontinue feeding during illness and advice aboutthe need to increase the sick child’s fluid intake.The denominator comprises sick children withvalidated classifications who do not need urgentreferral.There are two groups of interest: (1) sick childrenaged zero through five months and (2) sickchildren aged six to 23 months. The former shouldbe counseled for continued and increased ondemandbreastfeeding during illness (no foods orliquids). The latter should be counseled forcontinued and increased breastfeeding on-demand(greater frequency and longer), age-appropriatefeeding recommendations, and general increase offluids (breastfeeding and other).Strengths and LimitationsThe indicator does not measure whether themother or caretaker complies with the advice orwhether the advice on continued feeding specifiesthe types of food and frequency of feedingrecommended for the child’s age. As with manyof the indicators in chapter six, observation bias isof concern if the data are collected through directobservations of sick child consultations. Althoughexit interviews of caretakers may be a more cost-Growth Monitoring and Nutrition321

effective method of data collection, some degreeof misreporting of the content of counseling couldoccur. If no changes are observed in this indicatorover time, it may be useful to report separately thetwo elements of this indicator: continued feedingand increased fluid intake during illness.Sample QuestionsSample questions from the SPA (dated March2004) are the following:Option 1: Observation of sick child consultationo Record whether a provider did any of thefollowing when counseling the caretaker:(1) Provide general information about feedingor breastfeeding the child even when not sick• Yes/No/Don’t know/NA(2) Tell the caretaker to give extra fluids tothe child during this sickness• Yes/No/Don’t know/NA(3) Tell the caretaker to continue feeding thechild during this sickness• Yes/No/Don’t know/NAOption 2: Exit interviewo What did the provider tell you about feeding[NAME] during this illness?• Give less than usual• Give same as usual• Give more than usual• Give nothing/not feed• Didn’t discuss• Don’t knowo What did the provider tell you about givingfluids (or breastmilk, if the child is breastfed)to [NAME] during this illness?• Give less than usual• Give same as usual• Give more than usual• Give nothing/not feed• Didn’t discuss• Don’t know322 Chapter 8

PROPOROPORTIONOF HOSPITOSPITALALS AND MATERNITTERNITY FACILITIESDESIGNESIGNATEDAS BABABY FRIENDLRIENDLYWHO-Recommended indicator forNewborwborn n HealthDefinitionThe proportion of hospitals and maternityfacilities that have been accredited as “BabyFriendly” according to the ten UNICEF/WHOcriteria related to breastfeeding and newborn care.Numerator: Number of hospitals andmaternity facilities accredited as “BabyFriendly.”Denominator: Total number of hospitals andmaternity facilities that handle deliveries.To be designated as “Baby Friendly,” the hospitalmust:• Have a written breastfeeding policy that isroutinely communicated to all health care staff;• Train all health-care staff in the skills necessaryto implement this policy;• Inform all pregnant women about the benefitsand management of breastfeeding;• Help mothers initiate breastfeeding within anhour of birth;• Show mothers how to breastfeed and how tomaintain lactation, even if they should beseparated from their infants;• Give newborn infants no food or drink otherthan breast milk, unless medically indicated;• Practice “rooming in” by allowing mothers andinfants to remain together 24 hours a day;• Encourage breastfeeding on demand;• Give no artificial teats, pacifiers, dummies, orsoothers to breastfeeding infants; andGrowth Monitoring and Nutrition• Foster the establishment of breastfeedingsupport groups and refer mothers to them ondischarge from the hospital or birthing center.Measurement ToolsUNICEF/WHO/Wellstart Baby FriendlyHospitals Initiative internal self-assessment andexternal evaluation instrumentsWhat It MeasuresThis indicator provides useful information on theavailability of baby-friendly services in a givencountry. The Baby Friendly Hospitals Initiative(BFHI) is a joint UNICEF/WHO/Wellstartinitiative aimed at increasing breastfeeding ratesand encouraging global standards for maternityservices in hospitals and maternities.How to Measure ItData requirements are the number of maternitiesmeeting BFHI criteria and the total number ofmaternities and hospitals. Facilities first conducta self-assessment; then independent assessorsappointed by the national BFHI committee orUNICEF country offices evaluate them accordingto the above criteria. These same bodies aggregateinformation on the numbers and proportions offacilities acquiring “Baby Friendly” status fornational and global reporting (WHO, UNICEF,and Wellstart International, 1999)Strengths and LimitationsThe number of facilities achieving “Baby Friendly”status may be presented more often than theproportion because of difficulties in ascertainingthe total number of maternities required for thedenominator. Ascertaining the number ofmaternities in the private sector is particularlydifficult, and in many cases, private facilities maynot be represented in national estimates. The323

number of facilities achieving “Baby Friendly”status is of limited use for regional and crosscountrycomparisons because it is clearly affectedby geographic size. For example, by December2000, 6312 hospitals in China (or 47% of alleligible facilities) had achieved “Baby Friendly”status compared to 232 (or 66% of all eligiblefacilities) in Kenya.Second, the listing of facilities that are recordedas “Baby Friendly” may be out of date becauseperiodic reaccreditation to maintain standards isvoluntary and depends on the interest andmotivation of each individual facility. The date ofacquiring “Baby Friendly” status and whetherreaccreditation has occurred are not routinelyrecorded.324Chapter 8

EXCLCLUSIVEBREASREASTFEEDINTFEEDING RATEPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of infants aged 0-5 months who wereexclusively breastfed in the last 24 hours.Numerator: Number of infants aged 0-5 months(less than 180 days) who were exclusivelybreastfed in the last 24 hours.Denominator: Total number of infants aged 0-5months (less than 180 days) surveyed.Exclusive breastfeeding is the practice of givingbreast milk only to the infant, with no other solidsor liquids, including water. Infants are, however,allowed to have drops of vitamins/minerals/medicines (WHO, 1991).Measurement ToolsPopulation-based surveys employingrepresentative samples (e.g., DHS, KPC) andprogram records of exclusive breastfeeding rate (totrack trends but not impact)What It MeasuresThis indicator gives an overall measure of thedegree to which women have adopted behaviorsconsistent with the recommendation that infantsaged of 0-5 months should be exclusivelybreastfed.* Relative to infants who are exclusivelybreastfed, infants not breastfed at all have at least14 times the risk of death due to diarrhea. Therisk is greatest in the first two months of life(Murray et al., 1997). Even the introduction ofherbal teas and water to infants who have beenexclusively breastfed increases the risks of diarrhealGrowth Monitoring and Nutritionmorbidity and death. UNICEF and WHOrecommend that all women breastfeed theirchildren exclusively for the first six months.How to Measure ItThe data requirements are the number of livinginfants under six months of age and a 24 hourrecall of all liquids and solid food consumed byliving infants less than six months of age.Respondents should be probed about the differenttypes of liquids the infant may have received,including water, juice, milk, formula, and otherliquids. Both the DHS country reports andnutrition reports present the exclusivebreastfeeding rate (EBR) for infants less than fourmonths of age. However, programs can obtainand calculate the EBR for infants less than sixmonths of age using DHS data.Strengths and LimitationsThis indicator should be interpreted as theproportion of infants who “are currently beingexclusively breastfed,” rather than the proportionthat have been exclusively breastfed since birth.The use of a 24-hour recall period causes the* The 2001 UN policy statement on HIV and infant feedingis as follows: “When replacement feeding is acceptable,feasible, affordable, sustainable, and safe, avoidance of allbreastfeeding by HIV-infected mothers is recommended.Otherwise, exclusive breastfeeding is recommended duringthe first months of life. To minimize HIV transmission risk,breastfeeding should be discontinued as soon as feasible,taking into account local circumstances, the individualwoman’s situation, and the risks of replacement feeding(including infections other than HIV and malnutrition).When HIV-infected mothers choose not to breastfeed frombirth or stop breastfeeding later, they should be providedwith specific guidance and support for at least the first twoyears of the child’s life to ensure adequate replacementfeeding. Programmes should strive to improve conditionsthat will make replacement feeding safer for HIV-infectedmothers and families” (WHO, 2001).325

indicator to slightly overestimate the percent ofexclusively breastfed infants because some infantswho are given other liquids irregularly may nothave received them in the 24 hours before thesurvey. WHO’s Indicators for Assessing Breast-Feeding Practices (WHO, 1991), WellstartInternational’s Tool Kit for Monitoring andEvaluating Breastfeeding Practices and Programs,and the DHS reports all calculate EBR using the24-hour recall method.326Chapter 8

TIMELIMELY INITIANITIATIONTION OF BREASREASTFEEDINTFEEDINGDefinitionProportion of infants less than 12 months of agewho were put to the breast within one hour ofdelivery (WHO, 1991).Numerator: Number of infants less than 12months of age who were put to the breastwithin one hour of delivery.Denominator: Total number of infants lessthan 12 months of age.Measurement ToolsPopulation-based surveys employingrepresentative samples (e.g., DHS, MICS, KPC).Facility records may also be used to track trendsin breastfeeding initiation among clients but notto measure the impact of interventions on womenwith infants in the population of the catchmentarea.What It MeasuresThis indicator measures whether mothers in thepopulation and/or in health facilities initiate earlybreastfeeding with its respective benefits to bothmother (reduced postpartum hemorrhage) andinfant (skin-to-skin contact and exposure tomaternal antibodies in colostrum). Mothers aremore likely to successfully initiate lactation, toencounter fewer problems breastfeeding, and tomaintain optimal breastfeeding behaviors if theyinitiate breastfeeding shortly after birth.Breastfeeding should begin no later than one hourafter the delivery of the infant.of infants less than 12 months of age reported tohave been put to the breast within one hour ofbirth.When facility data are used to calculate thisindicator, the data requirements are the numberof infants discharged from the facility during thereference period and the number of infantsdischarged who were put to the breast within onehour of birth during the same reference period. Itis important to note that the two indicators(population-level and facility-level) are notcomparable.Strengths and LimitationsIn population-based surveys, mothers may havedifficulty recalling correctly when they initiatedbreastfeeding for their youngest children andwhether this was within one hour of delivery. Thisindicator may also mask changes in population orhealth facility practices that have occurred withinone year. The facility-based indicator does nothave as much recall bias. However, the facilitylevel indicator cannot be used to determinepopulation-level trends because it only reflectsbreastfeeding initiation by women who gave birthin facilities.Sample Questionso Did you ever breastfeed [NAME]?o How long after birth did you first put[NAME] to the breast?How to Measure ItThe data requirements for calculating thisindicator from population-based data are thefollowing: the number of infants less than 12months of age in the population and the numberGrowth Monitoring and Nutrition327

COMPLEMENTOMPLEMENTARARY FEEDINEEDING RATEPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of infants aged 6-9 months receivingbreastmilk and complementary food (based on last24 hours).Numerator: Number of infants aged 6-9months who received breastmilk andcomplementary foods in the last 24 hours.Denominator: Total number of infants aged6-9 months surveyed.Complementary foods are defined as solid or semisolid/mushyfoods; complementary foods do notinclude fluids.Measurement Tools:Population-based surveys employingrepresentative samples (e.g., DHS, KPC, MICS)What It MeasuresThe complementary feeding indicator is a basic,simple indicator that measures continuedbreastfeeding and consumption of complementaryfoods among children in the age group six throughnine months. Breast milk alone does not provideall the nutrients needed by an infant over sixmonths of age (Scrimshaw, 1996). UNICEF andWHO recommend that all women breastfeed theirchildren exclusively for the first six months. Afterthis age, the introduction of complementary foodsis critical to meet the protein, energy, andmicronutrient needs of the child. Continuing tobreastfeed with complementary feeding is alsoimportant (Dewey et al., 1996) as breastfeedingaccounts for a substantial proportion of fat, vitaminA, calcium, and quality protein into the secondyear of life (Murray et al., 1997).How to Measure ItThis basic calculation of complementary feedingrequires a sample of children 6-9 months of ageand information about feeding practices in the last24 hours, including breastfeeding status andwhether the child was given complementary foods.Facility records may be used to track trends incomplementary feeding but not to measure impact.Strengths and LimitationsThis indicator is simple to calculate and allowsfor a comparison of feeding practices for differentpopulation subgroups and an assessment ofchanges in feeding practices. The indicator canbe calculated for subgroups of children andparticipants in specific programs (e.g., programsthat promote good feeding practices among youngchildren and children seen in well-baby andimmunization clinics).However, this indicator has several limitations.First, it reflects only the prevalence ofcomplementary feeding. It does not allow one toassess the quality of food (energy density,micronutrient composition, or food handling),food quantity, or frequency of feeding. Second, itprovides minimal information on the extent towhich children are fed according to prescribedguidelines.Growth Monitoring and Nutrition329

MEANDIETIETARARY DIVERSITIVERSITY OF FOODSCONSUMEDBYCHILDRENAGEDGED6-23 MONTHSGED 6-23 MDefinitionMean number of food groups eaten in the last 24hours by children six through 23 months of age.Numerator: Sum of the number of food groups(0 through 8) eaten in the last 24 hours bychildren aged 6 through 23 months.Denominator: Total number of children aged6 through 23 months.“Dietary diversity” is defined here as the numberof food groups (0 through 8) eaten in the last 24hours.Measurement ToolsPopulation based surveys with 24-hour dietaryrecall information; DHS; KPC 2004What It MeasuresDietary diversity has long been recognized as akey element of high quality diets. A dietarydiversity indicator is based on the idea that morediverse diets are more likely than those that areless diverse to provide an adequate range ofnutrients. There is considerable evidence tosupport this idea (Ruel, 2003). Results fromseveral studies to date demonstrate that simpleindicators of dietary diversity (including meandietary diversity in children six through 23months, based on 24-hour food group recallquestions) reflect important nutritional differencesin child diet patterns and are positively associatedwith child nutritional status (Arimond and Ruel,2004).Dietary diversity is included in the GuidingPrinciples for Complementary Feeding of theBreastfed Child (PAHO/WHO, 2003), whereGuiding Principle #8 states, “Feed a variety offoods to ensure that nutrient needs are met. Meat,poultry, fish, or eggs should be eaten daily, or asoften as possible. Vegetarian diets cannot meetnutrient requirements at this age (six though 23months) unless nutrient supplements or fortifiedproducts are used. Vitamin A-rich fruits andvegetables should be eaten daily. Provide diets withadequate fat content. Avoid giving drinks withlow nutrient value, such as tea, coffee, and sugarydrinks such as soda. Limit the amount of juiceoffered so as to avoid displacing more nutrientrichfood.”How to Measure ItTo measure this indicator, the following foodgroup categories are summed, with each of thefood group categories scoring “1” if the child hadan item from that food group category yesterday,and “0” if the child did not. This results in a dietarydiversity score ranging from “0” to “8,” for eachchild. Complementary feeding questions arepresented in Annex 8.1 on page 369, and foodgroups can be aggregated into the eight categoriesas shown below.The food group categories are:• Grains, roots, and tubers (includingcomplementary foods like porridge, rice cereal,and gruel);• Vitamin A-rich fruits and vegetables;• Other fruits and vegetables;• Meat, poultry, fish, and shellfish (and organmeat, where eaten);• Eggs;• Legumes and nuts;• Dairy; and• Foods cooked with fat or oil.Growth Monitoring and Nutrition331

Strengths and LimitationsData collection is relatively simple and meandietary diversity is an easy indicator to understand.Because the mean dietary diversity indicator sumsthe total number of food groups eaten, higherscores correspond to a higher likelihood that anyparticular food group has been consumed.While there is a consensus that greater dietarydiversity is good, a number of outstandingquestions remain. Research is underway to furtherclarify the relationship between dietary diversityand nutrient adequacy. Furthermore, there are stillinconsistencies in the definition and measurementof dietary diversity, but comparability amongindicator results should improve as the foodgroupings and reference periods become morestandardized.Mean dietary diversity in children aged six through23 months is a meaningful measure of foodconsumption patterns at the population level.However, a single 24-hour measure of dietarydiversity in individual children six though 23months is not sufficient to provide usefulinformation about an individual child’s foodconsumption; multiple observations would benecessary.No specific guideline is available at this time toindicate an adequate or high level of dietarydiversity. The following ideas, however, guide thescoring for levels of dietary diversity. Whenchildren receive only one food, it is extremely likelyto be a staple food. A diversity score of “2” allowsonly one additional food group, and therefore thechild’s diet cannot meet the PAHO/WHO (2003)Guidelines, which recommend animal sourcefoods (supplement/fortified foods) and vitamin Arich plant foods daily. Therefore, children eating0-2 foods groups are considered to have “low”diversity. In contrast, children eating five or morefood groups in the previous day are very likely toreceive a variety of nutrient-dense foods and areconsidered to have “high” diversity. The middlediversity group includes children eating 3-4 foodgroups the previous day.332Chapter 8

6-23 MPROPOROPORTIONOF CHILDRENAGED6-23 MONTHSWITHGOODYOUNOUNG CHILDFEEDINEEDING PRARACTICESDefinitionProportion of children aged 6-23 months scoring“6” on the young child feeding practices score.Numerator: Number of children aged 6-23months scoring “6” on the young child feedingpractices score.Denominator: Total number of children aged6-23 months surveyed.“Young child feeding practices score:” This index isconstructed by combining scores for continuedbreastfeeding, age-appropriate frequency of feeding,and dietary diversity, as indicated in Table 8.2 below(Arimond and Ruel, 2003).Measurement ToolsDHS; KPCWhat It MeasuresThis indicator is a summary measure of theprevalence of optimal or adequate young childfeeding practices across the age range 6-23 months.The transition from exclusive breastfeeding to familyfoods typically covers the period from six to 18-23months of age – a vulnerable period during whichmalnutrition starts in many infants and youngchildren. Worldwide about 30% of children underfive are stunted because of poor feeding practices andrepeated infections. Appropriate feeding practicesare associated with improved nutrition and physicalgrowth, reduced susceptibility to common childhoodinfections, and better resistance to cope with them.These improved health outcomes in young childrenhave long-lasting health effects throughout thelifespan, including increased performance andproductivity, and reduced risk of certain noncommunicablediseases (WHO/UNICEF, 1998).How to Measure ItData to measure the young child feeding practicesscore can be collected through household surveys byasking the child’s mother or the caretaker a series ofquestions about whether they gave their infant/youngchild various foods in the previous 24 hours. Inaddition to questions about breastfeeding and liquidsand solids that the child had yesterday, a question onthe number of times the child was given semi-solidfoods the previous day is also asked. Complementaryfeeding questions in the KPC are the same as thosein the DHS and are presented in Annex 8.1.Table 8.2. Components and scoring of a young child feeding practices score for children aged 6-23 monthsElement includedContinued breastfeedingFrequency of feeding(Number of feeds yesterday)Dietary diversity(Number of food groups yesterday)Source: Arimond and Ruel, 2003, Table 7, p. 28.Growth Monitoring and NutritionScoringNo = 0Yes = 26-8 months:None = 0One = 12+ = 2Low (0-2) = 0Middle (3-4) = 1High (5-8) = 29-23 months:0-1 = 02 = 13+ = 2333

The young child feeding practices score is constructedby combining scores for continued breastfeeding, ageappropriatefrequency of feeding, and dietarydiversity. Table 8.2 details how each elementcontributes to a score for the index.The scoring for continued breastfeeding isstraightforward: either the mother is stillbreastfeeding the child or she is not, so scores areassigned easily. In the case of frequency of feeding,there is a clear recommendation. This scoring reflectsthe idea that there is a continuum. For example,clearly feeding a 6-8 months old infant once or twiceis better than not feeding solids/semi-solids at all;the highest score (2) should be assigned to the agespecificrecommended number of feeds.In the case of dietary diversity, no specific guidelineis available at this time to indicate an adequate levelof dietary diversity. The following ideas guide thescoring for diversity. There is evidence that morediverse diets are more likely to provide adequate levelsof a range of nutrients. When children receive onlyone food, it is extremely likely to be a staple food. Adiversity score of “2” allows only one additional foodgroup, and therefore the child’s diet cannot meet thePAHO/WHO (2003) Guidelines, whichrecommend animal source foods (supplement/fortified foods) and vitamin A rich plant foods daily.Therefore, children eating 0-2 foods groups areconsidered to have “low” diversity. On the other side,children eating five or more food groups in theprevious day are very likely to receive a variety ofnutrient-dense foods and are considered to have“high” diversity. The middle diversity group includeschildren eating 3-4 food groups the previous day.The scoring provides equal weights (two points) toeach of the three components (breastfeeding,frequency of feeding, and dietary diversity). Childrenare counted in the numerator if they scored six pointson the young child feeding practices score. Allchildren with a 0-5 on the young child feedingpractices score would in turn receive a zero (no) ongood young child feeding practices.Strengths and LimitationsThis indicator is relatively new. Its developmentstems from the recent WHO-led consensus buildingprocess of developing consistent, internationallyaccepted guidelines for breastfeeding andcomplementary feeding, and a set of indicators toallow assessment of feeding practices in the contextof surveys. Previous work in this area was hamperedby the lack of international standards and consensus,both on recommended practices and onmeasurement tools and indicators, forcomplementary feeding (Piwoz, Huffman, andQuinn, 2003). Ten “Guiding Principles” coveringall aspects of complementary feeding for breastfedchildren were recently developed by PAHO/WHO(2003) and a technical meeting to discuss indicatorsof complementary feeding and their field testing andvalidation was convened in December 2002 (Ruel,Brown, and Caulfield, 2003).The advantage of this indicator is that it capturesthe multidimensional and age-specific nature of childfeeding practices. It is useful, therefore, for assessingthe overall effectiveness of programs aimed atimproving complementary feeding practices,particularly if interventions target the entire rangeof practices included in the measure. As feedingpractices are likely to cluster, both at one point intime and over time (mothers who engage in earlypositive practices may also engage in a variety of betterpractices in subsequent years), the composite natureof the index is useful (Arimond and Ruel, 2003).An additional advantage of this indicator is that subsamplesizes in the age range 6-23 months are farmore likely to yield adequately precise estimates thanthe complementary feeding rate, which covers ages6-9 months only.However, some dimensions of feeding practices,specifically, the quality, texture, and nutrient densityof complementary foods, are not addressed by thisindicator. Further work using data from manycontexts could help define the relationship betweenthe young child feeding practices score and actualnutrient intake.No numeric targets have been set for this indicator.334Chapter 8

PROPOROPORTIONOF HOUSEHOLDSWITH ACCESSTOESSENTIALHANDANDWASHINASHING SUPPLIESCore e Hygiene Improvement IndicatorDefinitionProportion of households with access to essentialhandwashing supplies.Numerator: Number of households that haveaccess to essential handwashing supplies.Denominator: Total number of householdssurveyed.“Access” means that all essential items forhandwashing are either present (visible at the timeof survey) or can be produced within one minute.A special place for handwashing may not be alwaysfeasible, but ideally one should be located in ornear the toilet facility or kitchen.“Essential handwashing supplies” include all of thefollowing:1. Water (stored in a separate container, otherthan in the washing device);2. Soap (or locally available cleansing agent): and3. Washing device allowing for unassistedhandwashing (tap, basin, bucket, sink, or tippytap)Clean drying materials such as towels are notessential, because air drying is an acceptablealternative.Disposal of wastewater after handwashing doesnot require specific measures, unlike wastewaterfrom cleaning up children’s stool. However, lettingwastewater from handwashing accumulate inpuddles should be avoided to keep surroundingsdry and to prevent mosquitoes from breeding.Measurement ToolsPopulation-based surveys such as KPC and DHS/Environmental Health (EH) ModuleWhat It MeasuresBasic handwashing is an important element of thecontrol of diarrheal disease (DD). Handwashingbehavior is strongly influenced by the presence andaccess to water as well as access to essentialhandwashing supplies. To be optimally effective,the handwashing place should be located in closeproximity to the toilet facility so that householdmembers can conveniently wash their hands afterdefecation, or to the place where cooking takesplace so that food preparers can wash their handseasily before preparing food. At a minimum, thehandwashing place should be inside the yard.How to Measure ItData for calculating this indicator are collectedduring a household interview. A question is askedto determine where household members usuallywash their hands. The interviewer then asks toexamine the site and notes whether the sitecontains a water supply (it is desirable but notessential that this is of the improved type, becauseeven handwashing with water unsafe for drinkingcan be effective), a device for containing water andrinsing hands, and a cleansing agent such as soap.These items can either be displayed or broughtout within one minute for the household to qualifyas having access to essential handwashing supplies.To calculate the indicator, divide the number ofhouseholds with access to all essentialhandwashing supplies by the total number ofhouseholds in the sample.Strengths and LimitationsThe indicator does not measure the use ofhandwashing supplies at appropriate times orGrowth Monitoring and Nutrition335

knowledge of appropriate handwashingtechniques. Ideally, actual handwashing practicesshould be observed, but this is often not practicalduring household surveys. Some surveys do notcollect data on all criteria required in the definitionof appropriate handwashing places. The currentversion of the core questionnaire of the DHS, forexample, does not ask about a handwashing placeand supplies; however the DHS/EH module doesassess all essential criteria plus whether there isclean material for hand-drying, which can be usedwhere relevant to calculate an additional indicator.It is important, therefore, that baseline and followupsurveys use exactly the same methodology tocalculate the indicator so that any measurementbiases would be systematic.* Priority questions are needed for estimating theindicator described. Supplemental questions areoptional and may be useful for calculatingadditional indicators.Sample QuestionsPriority questions:*o Can you show me where you usually wash yourhands and what you use to wash hands?o Observation only: Is there water? Interviewer:turn on tap and/or a check container and noteif water is present or brought in one minuteor less.o Observation only: Is there soap or detergentor locally used cleansing agent? Note if presentor brought in one minute or less.o Observation only: Is there a hand-washingdevice such as a tap, basin, bucket, sink, ortippy tap present or brought in one minute orless?Optional supplemental questions:*o Observation only: Does the washing deviceallow unassisted washing and rinsing of bothhands, for example, a tap, basin, bucket, sink,or tippy tap?o Observation only: Is there a towel or cloth todry hands? Note if present or brought in oneminute or less.o Observation only: Does the towel or clothappear to be clean?336Chapter 8

PROPOROPORTIONOF HOUSEHOLDSWHERETHE CARETARETAKERAKEROF THE YOUNOUNGESGEST CHILDUNDERFIVEREPOREPORTEDAPPRPPROPRIAOPRIATEHANDANDWASHINASHING BEHAEHAVIORCore e Hygiene Improvement IndicatorDefinitionProportion of households where the caretaker of theyoungest child under five years reported using soap forwashing hands within the past 24 hours at two or morecritical times (after defecation and one of the followingfour: after changing a young child; before preparingfood; before eating; or before feeding a child).Numerator: Number of households where thecaretaker of the youngest child under five yearsreported using soap for washing hands withinthe past 24 hours at two or more critical times(after defecation and one of the following four:after changing a young child; before preparingfood; before eating; or before feeding a child).Denominator: Total number of households withchildren under five years surveyed.The above indicator is based on the assumption thatin each household only one caretaker (i.e., thecaretaker of the youngest child) will be interviewed.Where programs decide to interview more than onecaretaker and assess reported handwashing behaviorby caretakers of all children under five, the indicatorshould be calculated separately for children 0-23months and children 24-59 months. In addition,the indicator definition, numerator, and denominatorshould be modified as follows:Definition: Proportion of caretakers of children aged0-23 months (or some appropriate age range underfive years) who report using soap for washing handswithin the past 24 hours at two or more critical times(after defecation and one of the following four: afterchanging a young child; before preparing food; beforeeating; or before feeding a child).Growth Monitoring and NutritionNumerator: Number of caretakers of childrenaged 0-23 months (or some appropriate agerange under five years) who report using soapfor washing hands within the past 24 hours attwo or more critical times (after defecation andone of the following four: after changing a youngchild; before preparing food; before eating; orbefore feeding a child)Denominator: Total number of caretakers ofchildren aged 0-23 months (or some appropriateage range under five years) surveyed.“Appropriate handwashing behavior” includes twodimensions: use of soap and critical times forhandwashing.Critical times for handwashing listed by WHO are:• After defecation;• After handling a child’s feces/cleaning babies’bottoms/changing a young child;• Before food preparation;• Before eating; and• Before feeding a child.Measurement ToolsPopulation-based surveys (KPC and DHS/Environmental Health Module)What It MeasuresEvidence from trials and observational studies showthat handwashing with soap reduces the risk ofdiarrheal disease by 30-50% (Curtis and Cairncross,2003). This indicator inquires about actual behavior,and not knowledge. In many instances, the behaviorof the actual caretaker of the child (which could bethe mother, a sibling, other family, or other help withwhom the child spends most of his/her time) andthat of the household member who prepares foodwould be most important. Handwashing with soap337

at two critical times, “after defecation” plus anothercritical time, is suggested as a minimum but programsmay chose to set higher targets if more frequenthandwashing seems achievable. Although ash, sand,and mud are mentioned in the literature as localalternatives, neither their acceptability as a cleansingagent nor their actual use on a significant scale hasbeen established. The use of soap is promotedcommonly, for example through public-privatepartnerships for handwashing.How to Measure ItIn a household survey, this indicator is measured byself-reporting of critical times for handwashing; rarelyby demonstration of handwashing technique. Dataon handwashing techniques are collected by askingwhether the caretaker has soap, has used it in thepast 24 hours for handwashing, and the occasionsduring which soap was used for this purpose. The24 hour recall period can be approximated byrespondents mentioning “today” or “yesterday.” Ifonly one caretaker is interviewed per household, allhouseholds with children under five years oldsurveyed are counted in the denominator, whetheror not they have soap. Where other locallyappropriate cleansing materials are common (seeindicator about handwashing supplies), this indicatorcan be calculated only for households that have soap.If more than one caretaker is interviewed perhousehold, all caretakers of children under five arecounted in the denominator.In past household surveys, caretakers were frequentlyasked to name the critical times for washing hands.The question had multiple answers and measuredknowledge. Interviewers were instructed not to readthe answers out loud, but to record only thosementioned spontaneously by the caretaker.Unfortunately, these knowledge questions had littlediscriminatory power. Therefore, the soap usequestions mentioned above are now recommended.Social marketing and health extension/educationprograms have shown that considerableimprovement in handwashing behavior can beachieved over time (Bateman et al., 1995; Whitefordet al., 1996). Targets aimed at increasing appropriate338handwashing by 50% over the baseline are realisticand attainable.Strengths and WeaknessesAppropriate handwashing behavior includes threedimensions: critical times, frequency, as well astechnique. However, handwashing frequency andtechnique are difficult and time-consuming to assess.Requesting a handwashing demonstration and directobservation of the handwashing technique would bedesirable, but may be unfeasible in most surveysbecause it requires extensive training of the observersand is intrusive, time-consuming, and expensive.Handwashing behavior is strongly influenced by thepresence or absence of a convenient source of water.Where water is scarce, people may resort increasinglyto using recycled water for handwashing. Wherepossible, the use of recycled water for handwashingshould be assessed during the interview. Sincedifferent methods can be used to collect data onhandwashing, it is important that baseline andfollow-up surveys use exactly the same methodologyto calculate the indicator so that any measurementbiases would be systematic.It is also important to recognize that this indicator isbased partly on self-reported behavior in the past 24hours and does not indicate whether appropriatehandwashing at critical times is practiced routinely.Note that some large-scale surveys as the ICHS donot collect data on handwashing.Sample Questionso Do you have soap?o Have you used soap today or did you use soapyesterday?o When you used soap today or yesterday, whatdid you use it for? If “for washing my or mychildren’s hands is mentioned,” probe what wasthe occasion, but do not read the answers.Note that these questions are not available in theDHS.Chapter 8

6-23 MSICICK CHILDAGED6-23 MONTHSIS OFFEREDINCREASEDFLUIDSAND CONTINUEDFEEDINEEDINGPriorioritity y Indicator for IMCI at theHousehold LevelelDefinitionProportion of children aged 6-23 months whowere sick during the past two weeks and who wereoffered increased fluids and continued feedingduring the illness.Numerator: Number of children aged 6-23months who were sick in the past two weeksand who were offered increased fluids andcontinued feeding during the illness.Denominator: Total number of children aged6-23 months surveyed who were sick in thepast two weeks.“Fluids” include breastmilk; oral rehydration salts(ORS); recommended home fluids (RHF); andwater. Home fluids may include soups, cerealgruels, yogurt-based drinks, unsweetened fruitjuice, green coconut water, weak tea, plain cleanwater, or homemade sugar-and-salt solutions.Note that homemade sugar-and-salt solution isnot recommended in some settings due to thedifficulty of getting the quantities right.Measurement ToolsPopulation-based surveys, such as DHS, KPC, andMICSWhat It MeasuresThis indicator measures the performance ofprograms aimed at improving home casemanagement of sick children. The restriction ofthe indicator to children aged 6-23 months makesit more consistent with the PAHO/WHO (2003)Guiding Principles for Complementary Feeding of theBreastfed Child and with recommendations ofGrowth Monitoring and Nutritionexclusive breastfeeding for children who are undersix months of age. Any illness in children is likelyto reduce caloric intake and increase children’ssusceptibility to malnutrition following each illnessepisode. Both increased fluid intake and continuedfeeding during illness are important to reduce thisnutritional impact.How to Measure ItData requirements for calculating this indicator are:(1) the number of children age 6-23 months whowere sick in the two weeks preceding the survey andwhose caretakers offered them increased fluids andcontinued feeding; and (2) the total number ofchildren age 6-23 months surveyed who were sickin the two weeks preceding the survey.Caretakers of children who were sick in the twoweeks preceding the survey are asked whether thechild was given an increased amount of fluids, thesame amount, somewhat less, or much less to drinkwhile s/he was sick. Questions about feedingpractices during illness ask whether the child wasoffered less than usual to eat, about the sameamount, or more than usual to eat during theillness, and if less, how much less.Note that in the DHS, questions about increasedfluids and continued feeding during illness arerestricted to children who had diarrhea in the pasttwo weeks.A target of 80% was set for 2000 by the WorldSummit for Children (WHO, 1999b).Strengths and LimitationsThe “increased fluids” component of the indicatordoes not capture how soon after the start of theillness episode children are offered increased fluids.The timing of the administration of increasedfluids is especially important in diarrhea cases, as339

early increased fluid intake can prevent manychildren from becoming dehydrated and facilitatecontinued feeding by restoring appetite. Also, theindicator does not measure the nutritional valueof food given to child during illness.The indicator can be disaggregated into anindicator of increased fluid intake during illnessand an indicator of continued feeding duringillness. These separate indicators can be used forthe purpose of tracking the individual componentsof feeding practices during illness, especially if alow value is obtained on the basic indicator.Sample Questionso During [NAME’S] illness, did s/he drinkmuch less, about the same, or more than usual?o During [NAME’S] illness, did s/he eat less,about the same, or more food than usual?340Chapter 8

PROPOROPORTIONOF CHILDRENLIVINIVING IN HOUSEHOLDSUSINSINGADEQUDEQUATELTELY IODIZEDSALALTDefinitionProportion of children aged 6-59 months who livein households using adequately iodized salt.Numerator: Number of children aged 6-59months who live in households with saltcontaining 15+ parts per million (ppm) ofiodine.Denominator: Total number of children aged6-59 months.“Adequately iodized salt” is defined as salt containing15+ ppm of iodine.Measurement ToolsPopulation-based surveys; the testing of householdsalt for iodine is part of the core questionnaire inthe DHS and in other surveys such as the MICS.What It MeasuresThis indicator is a proxy measure for theproportion of children (breastfeeding and nonbreastfeeding)who may be receiving adequateamounts of iodine. The purpose of this indicatoris to evaluate the availability of adequately iodizedsalt in a given population. Iodine deficiencydisorders (IDD) are prevalent throughout theworld. WHO, UNICEF, and ICCIDDrecommend that the daily intake of iodine shouldbe 90 micrograms/millionth of a gram (µg) forpreschool children aged 0-59 months. IDDinterventions often focus on women ofreproductive age because of their increased needfor iodine during pregnancy. Iodine deficiency inpregnancy may impair the development of thefetus, and may cause extreme and irreversiblemental and physical retardation known ascretinism and other harmful effects. The mostcritical period is from the second trimester ofpregnancy to the third year after birth (WHO,2001a).How to Measure ItDuring household surveys, the iodine content ofsalt is determined using rapid test kits. These aresmall bottles of 10-50 ml containing a stabilizedstarch-based solution. One drop of solution placedon salt containing iodine (in the form of potassiumiodate) produces a blue/purple coloration, whichindicates that iodate is present (WHO, 2001a).UNICEF’s goal is to achieve the sustainableelimination of iodine deficiency disorders by 2005.Strengths and LimitationsData for calculating this indicator are easy tocollect in household surveys. Rapid test kits canbe used in the field to give immediate results.However, the kit can only assess whether the saltis iodized. It cannot reliably determine the iodineconcentration. The iodine concentration is mostaccurately measured by titration (liberating iodinefrom salt and titrating the iodine with sodiumthiosulphate using starch as an external indicator).Although titration is the preferred method foraccurate testing of the concentration of iodine insalt, it is time-consuming and not recommendedfor routing process monitoring.Sample Questionso INTERVIEWER: Ask respondent for ateaspoonful of cooking salt. Test salt foriodine. Record PPM (parts per million).Growth Monitoring and Nutrition341

12-59 MPROPOROPORTIONOF CHILDREN12-59 MONTHSWHOWEREDEWEWORMEDIN THE PASAST SIXMONTHSDefinitionProportion of children aged 12-59 months whoreceived a recommended deworming drug in thepast six months.Numerator: Number of children aged 12-59months who received a recommendeddeworming drug in the past six months.Denominator: Number of children aged 12-59 months surveyed.The “recommended deworming drug” is the onerecommended by the National Drug PolicyProgram and may include Albendazole,Levamisole, Mebendazole, Praziquantel, Pyrantel,or other WHO-recommended drugs.Measurement ToolsDHS; MICSWhat It MeasuresThis indicator measures drug coverage for soiltransmittedhelminth infections among childrenwho are under five years of age. Infants and youngchildren make up a substantial proportion ofpeople infected with or at risk of infection fromparasitic infections, such as roundworm,hookworms, or whipworms, in developingcountries. The prevalence of soil-transmittedhelminths (STH) among children less than 24months ranges from 20-80% (Crompton,Montresor, Nesheim, and Savioli, 2003;Montresor, Awasthi, and Crompton, 2003; WHO,2002, 2003). Schistosomiasis infection levels arenormally highest in school-age children who alsotend to have more rapid and higher reinfectionrates than older children and adults. Parasiticdisease is a major contribution to the malnutritioninfectioncomplex. Studies have shown that dewormingis associated with a decrease in theintensity of infection, improved nutritional statusin the form of increased weight gain, and improvediron status in anemic children (Montresor et al.,2003; Crompton et al., 2003).How to Measure ItThe data needed to measure this indicator arecollected during a household survey. Mothers withchildren under the age of five years are askedwhether the child was given any of therecommended antihelminthic medication over thelast six months. The global target set by the FiftyfourthWorld Health Assembly in 2001 is toregularly treat at least 75% of all school-agedchildren at risk of illness from schistosomiasis andsoil transmitted helminths by 2010 (WHO,2001c). Deworming programs may apply thistarget of 75% antihelminthic drug coverage tochildren aged 12-59 months.Strengths and LimitationsCollecting information for calculating thisindicator is relatively easy and calculation of theindicator is fairly simple. However, this indicatoris likely to suffer from reporting and recall bias.In addition, the respondent might not beknowledgeable about deworming status, ifdeworming was conducted in his or her absence.Growth Monitoring and Nutrition343

PREVREVALENALENCEOF NIGHTBLINDNESSIN CHILDRENSupplemental Indicator for IMCI atthe Household LevelelDefinitionProportion of children aged 24-59 months who arereported to be night blind by their caretakers (in areaswith high vitamin A deficiency).Numerator: Number of children aged 24-59months who are reported to be night blindby their caretakers.Denominator: Total number of children aged24- 59 months surveyed.Measurement ToolsPopulation-based surveys (micronutrient surveys,vitamin A surveys)What It MeasuresThis indicator measures the prevalence of nightblindness in children. Night blindness is an earlymanifestation of vitamin A deficiency in childrenmarked by poor adjustment to dim light. WHOand the International Vitamin A Consultative Group(IVACG) have established that if night blindnessprevalence among young children (18-59 or 25-59months) is higher than one percent, vitamin Adeficiency (VAD) is a problem of public healthsignificance in the community (WHO, 1982;Sommer, 1995).How to Measure ItIn order to determine whether a child isdemonstrating signs of night blindness, a caretakeris asked a number of questions: whether the childhas any problems seeing during the daytime; whetherthe child has any problems seeing in the nighttime;whether the child’s problems with night-time visionis different from that of other children in thecommunity. In many societies, there is a local termto describe the problem of night blindness that canhave a high sensitivity for finding night blindness.In addition to the questions described above,caretakers are also asked a direct question on whetherthe child has night blindness, using local terms thatdescribe the symptoms. Focus group discussions maybe helpful in identifying local terms or descriptionsof symptoms for night blindness and their usefulnessfor identifying VAD should be validated.In analyzing the data, children reported to have nightblindness but who also experience vision problemsduring the day should be excluded from thenumerator. The cut-off of 24 months is used becausebelow this age, night blindness is often not noticeddue to the child’s limited mobility.Strengths and LimitationsData on night blindness are easy to obtain when alocal term exists for the condition. In countries witha low prevalence of VAD, it may be difficult to finda widely recognized local term for night blindnessand interviewers must be carefully trained toadequately describe the condition. Where theprevalence of VAD is low, large samples may benecessary to detect changes at the population level.It is also important to note that night blindness canbe sensitive to seasons.Nightblindness is easier to detect in adults than inchildren. It was recently recommended that maternalnight blindness be adopted as an indicator of vitaminA deficiency in the community as a whole (IVACG,2002; Ramakrishnan and Darnton-Hill, 2002).However, maternal night blindness cannot be usedfor the purposes of monitoring and evaluating childbasedVADD programs.Growth Monitoring and Nutrition345

AMIN A SVITITAMINAMINA SUPPLEMENTUPPLEMENTATIONTIONSupplemental Indicator for IMCI atthe Household LevelelDefinitionProportion of children aged 6-59 months of agewho received a high dose of vitamin A in the lastsix months (in countries where there is a vitaminA supplementation policy).Numerator: Number of children aged 6-59months who received a high dose of vitaminA in the last six months.Denominator: Total number of children aged6-59 months surveyed.Doses are set by the national MOH or byfollowing WHO guidelines. Based on suggestionsfrom the IVACG meeting in 2000 in Annecy,France, and at the pre-XX IVACG meetingFebruary 11, 2001, in Hanoi Vietnam, therecommended schedule for routine high-dosevitamin A supplementation has been revised asoutlined in the table below (IVACG, 2002;Ramakrishnan and Darnton-Hill, 2002). Pleasenote that WHO has not yet adopted the guidelinesin the Annecy Accords:Measurement ToolsProgram statistics; vitamin A surveys; populationbasedsurveys such as DHS, KPC, and MICSWhat It MeasuresThis indicator measures the coverage achievedthrough national vitamin A supplementationprogram efforts in a specified period. Vitamin Adeficiency (VAD) is a major public health problemin developing countries. WHO estimates thatbetween 100 and 140 million children are vitaminA deficient. For children, lack of vitamin A causesvisual impairment, blindness, and significantlyincreases the risk of severe illness and death fromcommon childhood infections such as diarrhealdisease and measles. Supplementation as a vitaminA deficiency control strategy is the mostimmediate and direct approach to improvingvitamin A status and the one most widelyimplemented.How to Measure ItData requirements for calculating this indicatorare the number of children aged 6-59 monthssurveyed and the number of children aged 6-59months who received a high dose of vitamin A inthe past six months. In household surveys, it isInfants 0-5 monthsInfants 6-11 monthsChildren 12 months or older150,000 international units (IU)as three doses of 50,000 IU withat least a one-month intervalbetween doses100,000 IU as a single dose every4-6 months200,000 IU as a single dose every4-6 months.At each DTP contact (6, 10,and 14 weeks); otherwise atother opportunitiesAt every opportunity (e.g.,measles immunization)At any opportunityGrowth Monitoring and Nutrition347

est to ask the caretaker to see the child’simmunization card so as to ascertain the vitaminA supplementation status of the child and the datethe last dose of vitamin A was received. This willincrease the validity of survey information.In DHS surveys, women are shown the vitaminA capsule when collecting information aboutchildren’s vaccination status. This is because oralvitamin A supplements can be given to childrenin the form of liquid drops that come in capsulesor syrup and it is important to ensure that therespondent is clear that the questions of interestrefer to vitamin A as opposed to oral polio vaccine(OPV), which is also given as drops. If the childhas ever received a dose of vitamin A, the mother/caretaker is asked how many weeks or months agothe child received his/her last dose of vitamin A.This indicator can also be calculated from servicestatistics and program records. If the indicator isbased on an overall figure for the district, it isgenerally more accurate than if it is based on datafrom specific clinics. When the indicator is basedon service statistics, the numerator is the numberof children 6-59 months of age who received ahigh dose of vitamin A in the past six months andthe denominator is the number of children aged6-59 months in the target or service area. Thedenominators for the calculation of the vitamin Asupplementation rate for children aged 6 to 59months can be extrapolated from census data onthe age and sex distribution of the total population,using population projection models such asSPECTRUM (www.tfgi.com). When servicestatistics and program records are used, it isessential to specify whether this indicator measuressupplements distributed through outreach workersor only those given at fixed facilities, or both.UNICEF’s short-term goal (by 2005) is to doublethe number of countries with more than 70% vitaminA supplementation of children between the ages ofsix and 59 months. The UN Special Session onChildren in 2002 set as one of its goals theelimination of VAD and its consequences by 2010.Strengths and LimitationsThis indicator can be adapted to evaluate thedistribution of vitamin A capsules during NationalImmunization Days and National Vitamin AWeeks. If sample sizes permit, the indicator canalso be disaggregated by subgroups of thepopulation (urban/rural residence andsocioeconomic level) to examine whether theprogram is reaching certain target groups.However, this indicator is a coverage indicator anddoes not provide any information regarding theprevalence of vitamin A deficiency (as manifestedby night blindness, bitot spots, and cornealscarring). Although oral supplementation is usedfor both treatment and prevention of vitamin Adeficiency, it is not recommended as the only longtermapproach. In the home, vitamin A deficiencycan be prevented by the regular consumption ofvitamin A-rich foods, including fortified foods.Service statistics are relatively inexpensive tocollect and can be obtained at more frequentintervals than surveys. However, they are generallynot representative of an entire population. Sincethe quality of health statistics can vary amongfacilities, indicators calculated from servicestatistics may be less accurate than those based onsurvey data in places where the quality of routinedata is poor. In addition, it may be difficult toestimate the denominator for indicators based onservice statistics. The population denominatorsare often extrapolated from census data that areseveral years old. If population growth and ruralurbanmigration patterns have substantiallychanged over time, then census information maybe unsuitable for providing appropriatedenominators for local program managers todetermine vitamin A supplementation coverage.An alternative indicator reflecting the adequacyof the program in meeting children’s vitamin Aneeds is:• Number of capsules distributed per eligiblechild (age six through 59 months) in aspecified time period348Chapter 8

A six-month time frame is suggested for the abovealternative indicator. Note that some vitamin Asupplementation programs are timed to coincidewith low seasonal availability of vitamin A or highseasonal prevalence of illness. This would have tobe taken into account when using this alternativeindicator.Sample questionso Has [NAME] ever received a vitamin A doselike (this/any of these)?SHOW COMMON TYPES OF AMPULES/CAPSULES/SYRUPS.Growth Monitoring and Nutrition349

A D(SVITITAMINA DEFICIENEFICIENCYCY (SERERUMRETINETINOLCONONCENTRACENTRATIONTION)IVACG G Outcome IndicatorDefinitionProportion of children aged 6-59 months whoseserum vitamin A (retinol) concentration is lessthan 0.70 µmol/L.Numerator: Number of children aged 6-59months whose serum retinol concentration isless than 0.70 µmol/L.Denominator: Total number of children aged6-59 months surveyed.Children with serum retinol concentration lessthan 0.70 µmol/L are considered to be vitamin Adeficient.Measurement ToolsPopulation-based surveys that include thecollection of blood samples and measurement ofserum retinol (e.g. DHS)What It MeasuresThis indicator measures the prevalence of vitaminA deficiency (VAD) among children. It can beused by the MOH to determine the need forprograms to reduce levels of VAD, to help designand ensure they are targeted at areas of highestrisk, and also to monitor and evaluate their success.Originally, VAD was associated with loss of visionor with other vision impairments, including nightblindness, but research has now demonstrated adirect link between vitamin A deficiency andelevated mortality in children 6 to 59 months ofage. That link is due, in part, to the role played byvitamin A in the development of the immunesystem in young children.How to Measure ItLevels of serum retinol concentrations in bloodsamples are used to calculate this indicator. Thegold standard for measuring retinol concentrationsin blood is the drawing of venous blood for analysisusing high-performance liquid chromatography(HPLC). Alternative methods under developmentinclude a technique applied in the DHS, in whichblood spots from the finger or heel prick used foranemia testing are collected on a filter paper cardand the filter paper specimens are stored in aspecially designed box where they are protectedfrom sunlight and moisture while dryingovernight. The samples are conveyed to NorthCarolina to the only laboratory currently capableof analyzing the blood spots; the Program forAppropriate Technologies in Health (PATH) isdeveloping a technique that measures the RetinolBinding Protein (RBP) using an enzyme-linkedimmunosorbent assay (ELISA) machine. Neitherthe blood-spot technique nor the PATH test isfully validated.VAD is considered a public health problem if theprevalence of low serum retinol (

The practicality of this indicator is limited becausethe ease of collecting blood samples varies bysetting. It can be especially difficult to collectblood samples in populations with high prevalenceof HIV. There are also logistical difficulties inmaintaining the sample under the right storageconditions. In addition, it has been shown thatvitamin A levels derived from dried blood spotsamples are affected by the fact that the retinolbindingprotein in the serum collected on filterpaper decays in the first 7-10 days after collection.To account for this problem, the DHS collectsboth filter paper and venous blood samples for alimited number of children. Then, vitamin Alevels obtained from the survey are adjusted by arecovery factor, which is calculated as thecorrelation between the vitamin A levels measuredfrom dried blood spot retinol and plasma retinol.Another issue pertains to the validity of theindicator. The relationship between serum retinoland vitamin A status as indicated by body reservesis complex. Vitamin A circulates in the blood asretinol bound to its specific carrier protein, retinolbindingprotein (RBP). The level of retinol inthe blood is under homeostatic control over abroad range of body stores and reflects body storesonly when they are very low or very high. Sinceretinol-binding protein is an acute phase protein,acute and chronic infections can makeinterpretation of serum retinol levels difficult.Lack of laboratories for analysis of blood samplesis another limitation for this assessment.Finally, this indicator is not particularly useful forevaluating the most common vitamin Aintervention, the distribution of high-dose vitaminA supplements to children six to 59 months ofage. The supplement elevates the retinol levels inchildren for three to four months. Measurementstaken near the time of administration of thesupplements are artificially high whilemeasurements taken more than four to fivemonths after the administration of thesupplements may fail to capture the benefitschildren derive from having received thesupplement.352Chapter 8

HILDREN 6-59 MPROPOROPORTIONOF CHILDREN6-59 MONTHSWITH ANEMIADefinitionProportion of children aged 6-59 months of agewho have a hemoglobin level below 110 g/l.Numerator: Number of children aged 6-59 monthswho have a hemoglobin level below 110 g/l.Denominator: Total number of children aged 6-59months with valid hemoglobin measurement.Measurement ToolsDHS; routine data collection from health centersWhat It MeasuresAnemia is the most common and widespreadnutritional deficiency in the world. Nearly 40%of children aged 0-4 in non-industrializedcountries are estimated to be iron deficient(WHO, 2001b). Approximately half of anemiaprevalence can be attributed to iron deficiency, butthere are settings in which malaria or helminthinfections may cause more than 50% of anemiaseen in children. It is important, therefore, todetermine and address the multiple causes ofanemia in a population in an integrated fashion.Anemia is associated with premature birth, lowbirth weight, infections, and elevated risk of death.Anemia has also been associated with later physicaland cognitive impairment, resulting in loweredschool performance (Gleason, 2002; Lozoff, 2000;Pollit, 1997; WHO, 2001b). Among childrenunder five years of age, the greatest prevalence ofanemia occurs during the second year of life dueto low iron content in the diet and rapid growthduring the first year.Anemia occurs when mobilizable iron stores arefully depleted and the supply of iron to tissues iscompromised. When individual hemoglobinlevels are below two-standard deviation (-2 SD)of the distribution mean for hemoglobin in anotherwise normal population of the same genderand age who are living at the same altitude, theniron deficiency is considered to be present. In anormal population, 2.5% of the population wouldbe expected to be below this threshold. Anemiais considered to be a public health problem whenthe prevalence of concentration exceeds 5.0% ofthe population (WHO, 2001b).How to Measure ItThe data needed to measure this indicator arecollected during a household survey. Therecommended method for determininghemoglobin concentration in field surveys is theHemoCue system. It consists of a set of disposablecuvettes in which blood is collected and a batteryoperated photometer. The system is uniquelysuited for rapid field surveys because the one-stepblood collection and hemoglobin determinationdo not require addition of liquid reagent. Basedon prevalence estimated from blood levels ofhemoglobin, anemia in a population can becategorized into mild (5.0-19.9 %), moderate (20-39.9 %) and severe (40% or above).UNICEF’s goal is to reduce the prevalence ofanemia (including iron deficiency) by one thirdby 2010.Strengths and LimitationsData for calculating this indicator can easily becollected in household surveys using theHemoCue system, which has proven to be reliablein terms of precision against standard laboratorymethods. The indicator can be used to determinethe magnitude, severity, and distribution of irondeficiency and anemia among children who areunder five years of age in a population. This canprovide a basis for planning policies andGrowth Monitoring and Nutrition353

interventions or for identifying subgroups in whichinfants and young children are more affected or ata greater risk of iron deficiency and anemia. Theindicator can also be used to monitor and evaluatethe impact of interventions or provide a basis foradvocacy programs for iron deficiency and anemiaprevention.A major limitation of the indicator is that anemiais not a specific indicator of iron deficiency. Notall cases of anemia are due to iron deficiency andnot all iron deficiency will be reflected in anemia(Yip and Ramakrishnan, 2002). Other nutritionaldeficiencies (e.g., low intakes of folic acid andvitamins A, B 12and C) and infectious diseases (e.g.,malaria, schistosomiasis, and hookworm) may alsocontribute to anemia.354Chapter 8

OR-H-H/L(WLOW WEIGHTEIGHT-FOROR-HEIGHTEIGHT/L/LENENGTH(WASASTINTING)Supplemental Indicator for IMCI atthe Household LevelelDefinitionProportion of children (target age group) who arebelow –2 standard deviations from the medianweight-for-height according to the NCHS/WHOreference population.Numerator: Number of children (target agegroup) who are below –2 standard deviationsfrom the median weight-for-height accordingto the NCHS/WHO reference population.Denominator: Total number of children (targetage group) surveyed.The National Center for Health Statistics/WorldHealth Organization (NCHS/WHO) referencepopulation for weight-for-height provides standardsfor individual-level screening for malnutrition andpopulation-level monitoring. It is one of the mostcommonly used reference populations.The target age is that set by the national MOH policy.It may be set at under five years, under three years,or under two years of age.Measurement ToolsGrowth monitoring records; DHS; KPC; andhousehold nutrition surveysWhat It MeasuresWeight-for-height is an index that reflects bodyweight relative to height. Low weight-for-heighthelps to identify children suffering from current oracute undernutrition or wasting. Wasting is the resultof a weight falling significantly below the weightexpected of a child of the same length or height.Wasting indicates current or acute malnutritionresulting from failure to gain weight or actual weightloss. Causes include inadequate food intake, incorrectfeeding practices, diseases, and infections or, morefrequently, a combination of these factors (Cogill,2003; WHO, 1995). In humanitarian assistanceactivities, wasting or thinness in children aged 6-59months, combined with nutritional edema, is anindicator of acute malnutrition and should be usedto reflect the overall severity of a crisis. In general,weight-for-length (in children under two years ofage) or weight-for-height (in children over two yearsof age) is appropriate for examining short-termchanges, such as seasonal changes in food supply orshort-term nutritional stress brought about by illness.How to Measure ItWeight-for-height is calculated as the weight of eachchild in relation to the weight of a well-nourishedchild of the same sex and stature using the NCHS/WHO reference population. The data may becollected in household surveys such as the DHS andother nutrition surveys. Regular weighing andcharting of weight-for-height on the child’s growthchart provides another source of data for calculatingthis indicator.Children under two years of age are measured lyingdown on the board (recumbent length). Standingheight measurement is done for older children andis often referred to as stature. The child’s length orheight is recorded to the nearest 0.1 centimeter. Thechild’s measurement is then compared with themedian or average measurement for children of thesame age and sex in the NCHS/WHO standardreference population. More information on lengthand height measurement and the NCHS/WHOinternational reference tables can be found onFANTA’s Web site at www.fantaproject.org.Growth Monitoring and Nutrition355

Children’s weight may be measured by using a Salterlikehanging scale or a UNICEF mother/childelectronic scale, also referred to as the UNISCALE.If a UNISCALE is used, the enumerator first recordsthe mother’s weight while she is holding the baby.The weight is recorded in kilograms to one decimalpoint. A second weight reading is then recordedwith just the mother standing on the scale. Thedifference in the readings is the weight of the child.The standard deviation unit or Z-score is the simplestway of making comparisons to the referencepopulation. The Z-score is defined as the differencebetween the value for an individual and the medianvalue of the reference population in the same weightor height, divided by the standard deviation of thereference population. The median is the value atexactly the mid-point between the largest andsmallest.The cut-off points for different malnutritionclassifications under the NCHS/WHO system islisted below:356Between –1 SD and –2 SDBetween –2 SD and –3 SDBelow –3 SDmildmoderatesevereChildren who are below -2 SD from the median ofthe NCHS reference population in terms of weightfor-heightare considered thin or wasted. Inhumanitarian assistance activities, it is recommendedto include all children under the age of five yearswho have pitted edema in their limbs in thenumerator, irrespective of their anthropometricstatus. This is because edema is a clinical sign ofsevere malnutrition and shows strong associationswith mortality.In clinic-based growth monitoring, the Road-to-Health (RTH) system is typically used instead ofthe NCHS/WHO system. When examiningmalnutrition patterns over time, it is important touse the same system to analyze and present trenddata because the cutoff points for mild, moderate,and severe malnutrition are different in eachclassification system.In non-disaster areas, the prevalence of low weightfor-heightis usually at less than 5% (WHO, 1995).Prevalence of wasting between 10-14 % is consideredserious and prevalence greater than 15 % is consideredcritical.Strengths and LimitationsThis indicator is simple to calculate and is usefulwhen exact ages are difficult to determine. Lowweight-for-height can be used as a screening ortargeting indicator, for example to identify infants/children who need supplementary or therapeuticfood and/or treatment for diseases, particularlydiarrhea. In emergency settings, weight-for-heightis a useful indicator for screening and surveillance.As mentioned above, humanitarian activities alsoneed to take into account nutritional edema whencalculating this indicator. Percentage of the referencemedian should be reported as well, as it is used as anentry criterion for feeding programs. Becausewasting in individual children and population groupscan change rapidly, the indicator is responsive toshort-term program influences. However, theindicator is also highly susceptible to seasonalvariations in food availability. Weight-for-height isnot recommended for evaluating change inanthropometric status in non-emergency situations.The main limitation of this indicator is that weightand height can be difficult to obtain, leading toproblems of validity of measurement. The mostfrequent problems in height measurement areinadequate positioning of the child’s head and feet, areading done in an oblique position, and not facingthe reading point of the measuring board or heightmeasuringapparatus. The largest acceptabledifference between repeated measurements of heightis 1.0 cm. For weight, the largest acceptable differencebetween repeated measurements is 0.5 kg (Cogill,2003). If repeated measurements are different fromeach other, the measurements should be disregardedand the measuring should start again. Enumeratorvariability in weight measurement can be reducedthrough good training and supervision.Chapter 8

L(SLOW HEIGHTEIGHT/L/LENENGTHTH-FOROR-AGE(STUNTINTUNTING)Supplemental Indicator for IMCI atthe Household LevelelDefinitionProportion of children (target age group) surveyedwho are below –2 standard deviations from themedian height/length-for-age of the NCHS/WHO reference population.Numerator: Number of children (target agegroup) surveyed who are below –2 standarddeviations from the median height/lengthfor-ageof the NCHS/WHO referencepopulation.Denominator: Number of children (target agegroup) surveyed.The National Center for Health Statistics/WorldHealth Organization (NCHS/WHO) referencepopulation for height-for-age provides standardsfor individual-level screening for malnutrition andpopulation-level monitoring. It is one of the mostcommonly used reference populations.The target age is that set by the national MOHpolicy. It may be set at under five years, underthree years, or under two years of age.Measurement ToolsGrowth monitoring records; DHS; KPC; andhousehold nutrition surveysGrowth Monitoring and NutritionWhat It MeasuresLow height-for-age reflects past undernutritionor a past growth failure and is a measure of stuntingin an individual or the extent of stunting in apopulation. For children below two years of age,the term is length-for-age. Above two years ofage, the term is height-for-age. Low length-foragestemming from a slowing in the growth ofthe fetus and the child and resulting in a failure toachieve length as compared to a healthy wellnourishedchild of the same age is a sign ofstunting. Low height-for-age is associated with anumber of long-term factors including chronicinsufficient food intake, frequent infection,sustained inappropriate feeding practices, andpoverty. Worldwide about 30% of children underfive are stunted because of poor feeding practicesand repeated infections.In areas of high prevalence, the age of the childmodifies the interpretation of height-for-age. Foryounger children (under 24-35 months), lowheight-for-age reflects a continuing process of“failing to grow” or stunting. For older children,it reflects the state of “having failed to grow” orbeing stunted (WHO, 1995). For these reasons,the prevalence of stunting disaggregated by agecategories and sex is easier to interpret.How to Measure ItTo measure this indicator, data are collected onthe height and date of birth for all children, usuallyunder five years of age. The child’s age incompleted months is required for sampling. Tomeasure the age of the child, enumerators need toexamine documentary evidence of the birth date(such as a birth or baptismal certificate, clinicalrecords, etc.) in addition to asking the mother, inorder to minimize recall errors. If the child’s dateof birth cannot be recalled, a local calendar shouldbe used to assist the mother in recalling the dateof birth.The age of the child determines whether the childis measured standing or lying down for height orlength, and for converting height into standardindices. Children under two years of age aremeasured lying down on the board (recumbent357

length). Standing height measurement is donefor older children and is often referred to as stature.The child’s length or height is recorded incentimeters up to one decimal point. The child’smeasurement is then compared with the medianor average measurement for children of the sameage and sex in the NCHS/WHO standardreference population. More information oncollection of weight and height data and theNCHS/WHO international reference tables canbe found on the Food and Nutrition TechnicalAssistance (FANTA) project’s Web site atwww.fantaproject.org.The standard deviation unit or Z-score is thesimplest way of making comparisons to thereference population. The Z-score is defined asthe difference between the value for an individualand the median value of the reference populationin the same age or height, divided by the standarddeviation of the reference population. The medianis the value at exactly the mid-point between thelargest and smallest.The cut-off points for different malnutritionclassifications under the NCHS/WHO system islisted below:Between –1 SD and –2 SDBetween –2 SD and –3 SDBelow –3 SDmildmoderatesevereChildren who are below –2 SD from the medianof the NCHS reference population in terms ofheight-for-age are considered short for their ageor stunted.In clinic-based growth monitoring, the Road-to-Health (RTH) system is typically used instead ofthe NCHS/WHO system. When examiningmalnutrition patterns over time, it is important touse the same system to analyze and present databecause the cutoff points for mild, moderate, andsevere malnutrition are different in eachclassification system.Data on the prevalence of stunting should beexamined separately by age categories and sex. Inareas of high prevalence, the age of the childmodifies the interpretation of height-for-age. Foryounger children (under 24-35 months), lowheight-for-age reflects a continuing process of“failing to grow” or stunting. For older children,it reflects the state of “having failed to grow” orbeing stunted (WHO, 1995).One of the goals established at the May 2002United Nations General Assembly Special Sessionon Children is the reduction of malnutritionamong children under five years of age by at leastone third, with special attention to children undertwo years of age (United Nations, 2002). It isrecommended not to measure height-for-ageannually as the prevalence of growth failure orstunting would not be expected to change in thattime frame.Strengths and LimitationsStunting is a measure of the long-term effects ofmalnutrition in a population and is unaffected byseasonal variation. It provides a better indicationof trends than the wasting indicator (low weightfor-height),since it reflects long-term outcomes,such as frequent and high disease burden, limitedaccess to food supply, poor feeding practices, and/or low household socioeconomic status, in thetarget population. Because stunting in childrenreflects socioeconomic conditions that are notconducive to good health and nutrition, thisindicator is often used to target developmentprograms. A decrease in the prevalence of stuntingat the population level is a long-term indicatorthat social development is benefiting the poor aswell as the relatively wealthy. Information onstunting for individual children is also usefulclinically as an aid to diagnosis. Stunting basedon height-for-age can be used for evaluationpurposes but it is not recommended formonitoring as it does not change in the short term,such as 6-12 months.358Chapter 8

Some programs report the prevalence of stuntingfor children under age 24 months rather than therecommended 24-59 months. Restricting the agegrouping to children under age 24 months has thedisadvantage of not capturing the lagged effect ofthe program and reducing the number of potentialparticipants in a survey. The advantage of usingchildren under age 24 months is that theprevalence of stunting in children of this age groupis likely to be more responsive to the impact ofinterventions than in older children. The data forchildren under age 24 months may be more useful,therefore, for determining the factors related tostunting for program design and redesign (Cogill,2003).The main limitation of this indicator is that lengthor height can be a difficult to obtain, thus leadingto problems of validity. The most frequentproblems in height measurement are inadequatepositioning of the child’s head and feet, a readingdone in an oblique position, and not facing thereading point of the measuring board or heightmeasuringapparatus. The largest acceptabledifference between repeated measurements ofheight is 1.0 cm. If repeated measurements aredifferent from each other, the measurementsshould be disregarded and the measuring shouldstart again. Accuracy of measurement is achievedthrough good training and supervision.Growth Monitoring and Nutrition359

(ULOW WEIGHTEIGHT-FOROR-AGE(UNDERWEIGHTEIGHT)Millennium lennium Development elopment Goal(MDG) G) IndicatorDefinitionProportion of children (target age group) surveyedwho are below –2 standard deviations from themedian weight-for-age according to the NCHS/WHO reference population.Numerator: Number of children (target agegroup) surveyed who are below –2 standarddeviations from the median weight-for-age ofthe NCHS/WHO reference population.Denominator: Number of children (target agegroup) surveyed.The National Center for Health Statistics/WorldHealth Organization (NCHS/WHO) referencepopulation for weight-for-age provides standardsfor individual screening for malnutrition andpopulation-level monitoring. It is one of the mostcommonly used reference populations.The target age is that set by the national MOHpolicy. It may be set at under five years, underthree years, or under two years of age.Measurement ToolsGrowth monitoring records; DHS; KPC; andhousehold nutrition surveysWhat It MeasuresWeight-for-age reflects body mass relative tochronological age. Low weight-for-age identifiesthe condition of being light or underweight for aspecific age and reflects the process of gaininginsufficient weight relative to age or losing weight.Since weight-for-age is influenced by both theheight of the child and by its weight, the indicatorGrowth Monitoring and Nutritionreflects both past (chronic) and/or present (acute)undernutrition. This indicator is also a measureof health and nutritional risk in a population.Pelletier and colleagues (1995) calculated that therelative risk of death was 2.5, 4.6, and 8.4 for mild,moderate, and severe undernutrition (low weightfor-age),respectively. Thus, monitoring weightfor-agecan help assess the contribution of growthpromotion programs to mortality reduction.Underweight, based on weight-for-age, isrecommended as the indicator to assess changesin the magnitude of malnutrition over time.How to Measure ItTo measure this indicator, data are collected onthe weight and date of birth of the child for allchildren under five years of age. The child’s agein completed months is required for sampling andfor converting weight into standard indices. Tomeasure the age of the child, enumerators need toexamine documentary evidence of the birth date(such as a birth or baptismal certificate, clinicalrecords, etc.) in addition to asking the mother, inorder to minimize recall errors. If the child’s dateof birth cannot be recalled, a local calendar shouldbe used to assist the mother in recalling the dateof birth.Children’s weight may be measured by using aSalter-like hanging scale or a UNICEF mother/child electronic scale, also referred to as theUNISCALE. If a UNISCALE is used, theenumerator first records the mother’s weight whileshe is holding the baby. The weight is recorded inkilograms to one decimal point. A second weightreading is then recorded with just the motherstanding on the scale. The difference in thereadings is the weight of the child.The standard deviation unit or Z-score is thesimplest way of making comparisons to the361

eference population. The Z-score is defined asthe difference between the value for an individualand the median value of the reference populationin the same age or weight, divided by the standarddeviation of the reference population. The medianis the value at exactly the mid-point between thelargest and smallest.The cut-off points for different malnutritionclassifications under the NCHS/WHO system islisted below:Between –1 SD and –2 SDBetween –2 SD and –3 SDBelow –3 SDmildmoderatesevereChildren who are below –2 SD from the medianof the NCHS reference population in terms ofweight-for-age are considered underweight fortheir age.In clinic-based growth monitoring, the Road-to-Health (RTH) system is typically used instead ofthe NCHS/WHO system. When examiningmalnutrition patterns over time, it is important touse the same system to analyze and present databecause the cutoff points for mild, moderate, andsevere malnutrition are different in eachclassification system.Data on the prevalence of underweight should beexamined separately by age categories and sex.Underweight can be reported annually as it isinfluenced by short-term effects such as a recentoutbreak of diarrheal diseases. Household surveysare generally conducted every three to five years.Strengths and LimitationsWeight-for-age is influenced by both height ofthe child (height-for-age) and by its weight(weight-for-height). The advantage of thisindicator is that it reflects both present (acute) and/or past (chronic) undernutrition. Weight-for-ageis a useful tool in clinical settings for continuousassessment of nutritional progress and growth. Itcan be used at the individual level to identifyinfants and children with poor health andnutrition, for interventions specially tailored tocauses of poor growth (e.g., breastfeeding support,nutrition education, supplementation of infant/child and mother, prevention of diarrhea), and toassess response to the intervention(s). At theindividual level, it can also be used to identifyinfants and children who are failing to thrive andwho need treatment for underlying diseases andmonitoring their response to therapy. However,the composite nature of the index makesinterpretation difficult.One of the major limitations of the indicator isthe issue of the reliability of weight measurements.There may be some degree of variability betweeninterviewers in performing the task of weighing.For weight, the largest acceptable differencebetween repeated measurements is 0.5 kg (Cogill,2003). Enumerator variability in weightmeasurement can be reduced through extensivetraining. The validity of this indicator alsodepends on the accuracy of the weighinginstruments and the caretaker’s ability to reportthe correct age of the child.One to the goals established at the May 2002United Nations General Assembly Special Sessionon Children is the reduction of malnutritionamong children under five years of age by at leastone third, with special attention to children undertwo years of age (United Nations, 2002).362Chapter 8

-FUNDERNDER-F-FIVEMORORTALITALITY RATETETE (U5MR)Millennium lennium Development elopment Goal(MDG) G) IndicatorIndicator Cross-referenced in Chapter NineDefinitionThe under-five mortality rate (U5MR) is thenumber of deaths of children aged 0-4 years in aspecified year per 1000 live births in that specifiedyear.Numerator: Number of deaths of children aged0-4 years in a specified year x 1000.Denominator: Total number of live births inthat specified year.A “live birth” is described by the United Nations(1955) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”The U5MR may also be expressed as theprobability (expressed as a rate per 1,000 livebirths) of a child born in a specified year dyingbefore reaching the age of five if subject to currentage-specific mortality rates.Measurement ToolsCensus; population-based surveys (e.g., DHS,MICS); vital registrationWhat It MeasuresThis indicator measures the risk of dying ininfancy and early childhood. It also reflects thesocial, economic, and environmental conditions inwhich children (and others in society) live,including their health care. The under-fivemortality rate captures more than 90% of globalmortality among children under the age of 18. Theindicator may be used, therefore, as a measure ofchildren’s well-being and the level of effort beingmade to maintain child health.How to Measure ItAt the national level, the best source of data is acomplete vital statistics registration system – onecovering at least 90% of vital events in thepopulation. Such systems are uncommon indeveloping countries, so estimates are alsoobtained from sample surveys or derived byapplying direct and indirect estimation techniquesto registration, census, or survey data.A wide variety of household surveys, includingthe MICS and DHS, are used in developingcountries to estimate under-five mortality rates.In the DHS, birth histories are collected fromwomen aged 15-49 years. Each eligible woman isasked the dates of birth and, where relevant, theage of death of every live-born child she has had.Using life-table analysis, this approach producesdirect estimates of under-five mortality.Growth Monitoring and Nutrition363

A second survey method is called the Brassmethod, which asks each woman surveyed simplequestions such as her age, the total number ofchildren she has borne, and the number of thosechildren who have died. The age of the motherand the proportion of children that died, inconjunction with assumption of fertility and otherfactors, allow an estimation of under-five mortality.This approach is used by MICS and producesindirect estimates of under-five mortality.Longitudinal or prospective surveys following asample of babies born in a specified year over anumber of years may also be used to estimateunder-five mortality. However, this approach isrelatively uncommon in developing countries.Vital statistics are typically available once a year,but they are unreliable in most developingcountries. Household surveys that includequestions on births and deaths are generallyconducted every three to five years. Target 5 ofthe Millennium Development Goals (MDG) isto reduce the under-five mortality rate by twothirdsbetween 1990 and 2015.surveys it is important to consider sampling errors.In addition, indirect estimates rely on estimatedactuarial (“life”) tables that may be inappropriatefor the population concerned.The U5MR has been recommended as anindicator for determining the extent of vitamin Adeficiency (VAD) in a population. This is becausepopulations with high U5MR invariably tend tohave significant VAD. An U5MR > 50 per 1000live births is considered to indicate a highlikelihood of VAD problems at the country level(IVACG, 2002). Under-five mortality rates arealso frequently used to compare levels ofsocioeconomic development across countries.Comparisons between countries should be madewith caution as the data used to arrive at estimatesof under-five mortality come from a wide varietyof sources of disparate quality and may refer todifferent reporting periods.Strengths and LimitationsVital registration data may suffer from a numberof problems. These include the omission of somebirths and deaths, especially infants that diedshortly after birth, and the misreporting of thedate of birth and age at death. Omission of underfivedeaths is usually most severe for deaths thatoccur early in infancy.In developing countries, household surveys areessential to the calculation of this indicator, butthere are some limits to their quality. Survey dataare subject to recall error, and surveys estimatingunder-five deaths require large samples, whichmay be costly. Therefore, when using household364Chapter 8

RefererencesencesArimond M, and Ruel MT. (2003). Generating Indicators of Appropriate Feeding of Children 6 through 23Months from the KPC 2000+. Washington, DC: Food and Nutrition Technical Assistance (FANTA) Project,Academy for Educational Development.Arimond M, and Ruel MT. (2004). Dietary Diversity, Dietary Quality, and Child Nutritional Status: Evidencefrom Eleven Demographic and Health Surveys. Washington, DC.: Food and Nutrition Technical Assistance(FANTA) Project, Academy for Educational Development.Beaton G, Kelly A, Kevany J, Martorell R, and Mason J. (1990). Appropriate Uses of Anthropometric Indicesin Children. Geneva, Switzerland: ACC/SCN. ACC/SCN State of the Art Series, Nutrition PolicyDiscussion Paper No. 7.Bertrand JT, and Escudero G. (2003). Compendium of Indicators for Evaluating Reproductive Health Programs.Chapel Hill, NC: MEASURE Evaluation.Cogill B. (2003). Anthropometric Indicators Measurement Guide. Washington, DC: Food and NutritionTechnical Assistance (FANTA) Project, Academy for Educational Development.Crompton DWT, Montresor A, Nesheim MC, and Savioli L (eds.). (2003). Controlling Disease Due toHelminth Infections. Geneva, Switzerland: World Health Organization.Dewey KB, and Brown KH. (2003). Update on Technical Issues Concerning Complementary Feeding ofYoung Children in Developing Countries and Implications for Intervention Programs. Food and NutritionBulletin, 24(1):5-28.Galloway R. (2003). Anemia Prevention and Control: What Works, Part I Program Guidance and Part II Toolsand Resources. Washington, DC: USAID.Gleason GR. (2002). Iron Deficiency Anemia Finally Reaches the Global Stage of Public Health. Nutritionin Clinical Care, 5(5):217-219.Habicht JP, Mason J, and Tabatabai H. (1984). Basic Concepts for the Design of Evaluation DuringProgramme Implementation. In: Sahn D, Lockwood R, and Scrimshaw N (eds.). Methods for the Evaluationof the Impact of Food and Nutrition Programmes. UNU, Tokyo. 1 25.Habicht JP et al. (1999). Evaluation Designs for Adequacy, Plausibility and Probability of PublicHealth Programme Performance and Impact. International Journal of Epidemiology, 28:10-18.IVAGG. (2002). The Annecy Accords to Assess and Control Vitamin A Deficiency: Summary of Recommendationsand Clarifications. Washington, DC: ILSI Research Foundation.Lozoff B. (2000). Perinatal Iron Deficiency and the Developing Brain. Pediatric Research, 48:137-139.Mason J, Habicht JP, Tabatabai H, and Valverde V. (1984). Nutritional Surveillance. Geneva, Switzerland:World Health Organization. Monograph, 194 pp. ISBN 92 4 156078 9.Growth Monitoring and Nutrition365

Mason JB, and Habicht JP. (1984). Stages in the Evaluation of On-going Programs. In: Sahn D, LockwoodR, and Scrimshaw N (eds.). Methods for the Evaluation of the Impact of Food and Nutrition Programmes, p. 26-45. Tokyo: United Nations University.Mason JB, Mannar V, and Mock N. (1999). Controlling Micronutrient Deficiencies in Asia. AsianDevelopment Review, 17(1-2)66-95.Mason J, Smith N, Deitchler M, Wetzler E, Agyeman Y, and Dobbs D. (2000). CARE Food Review: AReview of CARE's Use of Title II Resources. (1) Overview (2) Country Case Studies (Bangladesh, Bolivia,Guatemala, Haiti, Honduras, India, Kenya, Madagascar, Mozambique, and Peru). Mimeo. New Orleans:Department of International Health and Development, Tulane School of Public Health and Tropical Medicine.Mason JB, Lotfi M, Dalmiya N, Sethuraman K, and Deitchler M, with Geibel S, Gillenwater K, Gilman A,Mason K, and Mock N. (2001a). The Micronutrient Report: Current Progress in the Control of Vitamin A,Iodine, and Iron Deficiencies. Ottawa, Canada: Micronutrient Initiative/International Development ResearchCenter.Mason JB, Hunt J, Parker D, and Jonsson U. (2001b.) Improving Child Nutrition in Asia. Supplement toFood and Nutrition Bulletin, 22(3)5-80.Montresor A, Awasthi S, Crompton DWT. (2003). Use of Benzimidazoles in Children Younger than 24Months for the Treatment of Soil-Transmitted Helminthiasis. Acta Tropica, 86(2-3):223-32PAHO/WHO. (2003). Guiding Principles for Complementary Feeding of the Breastfed Child. Washington,DC/Geneva, Switzerland: PAHO/WHO. Retrieved from http://www.who.int/child-adolescent-health/New_Publications/NUTRITION/guiding_principles.pdfPelletier D. (1994). The Relationship between Child Anthropometry and Mortality in Developing Countries:Implications for Policy, Programs and Future Research (2047S). Supplement to Journal of Nutrition, 124:10S.Pelletier DL, Frongillo Jr. EA, Schroeder DG, and Habicht JP. (1995). The Effects of Malnutrition onChild Mortality in Developing Countries. Bulletin of the World Health Organization, 73(4):443-8.Pinstrup Andersen P, Burger S, Habicht JP, and Peterson K. (1993). Protein Energy Malnutrition. In:Disease Control Priorities in Developing Countries. Oxford Medical Publications, 391-420.Piwoz EG, Huffman SL, and Quinn VJ. (2003). Promotion and Advocacy for Improved ComplementaryFeeding: Can We Apply the Lessons Learned from Breastfeeding? Food and Nutrition Bulletin, 24:29-44.Pollit E. (1997). Iron Deficiency and Educational Deficiency. Nutrition Review, 55:133-144.Ramakrishnan U, and Darnton-Hill I. (2002). Assessment and Control of Vitamin A Deficiency Disorders.Journal of Nutrition, 132:2947S-2953S.Ruel MT. (2003). Operationalizing Dietary Diversity: A Review of Measurement Issues and ResearchPriorities. Journal of Nutrition, 133:3911S-3926S.366Chapter 8

Ruel MT, Brown KH, and Caulfield LE. (2003). Moving Forward with Complementary Feeding: Indicatorsand Research Priorities. Washington, DC: International Food Policy Research Institute. Food Consumptionand Nutrition Division Discussion Paper No. 146.Sanders D. (1999). Success Factors in Community Based Nutrition Programmes. Food & Nutrition Bulletin,20(3):307-314.Sommer A. (1995). Vitamin A Deficiency and Its Consequences: A Field Guide to Detection and Control, 3rdEd. Geneva, Switzerland: World Health Organization.Swindale A, Deitchler M, Cogill B, and Marchione M. (2004). Impact of Title II Maternal and Child Healthand Nutrition Programs on the Nutritional Status of Children. Washington, DC: Food and Nutrition TechnicalAssistance Project, Academy for Educational Development. Retrieved from http://www.fantaproject.org/downloads/pdfs/ffpOP4.pdfTomkins AM, and Watson F. (1989). Malnutrition and Infection: A Review. Geneva, Switzerland: ACC/SCN and World Health Organization. ACC/SCN State of the Art Series, Nutrition Policy DiscussionPaper No. 5.Tontisirin K, and Winichagoon, P. (1999). Community Based Programs: Success Factors for Public NutritionDerived from Thailand's Experience. Food & Nutrition Bulletin, 20(3):315-322.UNICEF. (1990). Strategy for Improved Nutrition of Children and Women in Developing Countries. NewYork: UNICEF. Policy Review Paper E/ICEF/1990/1.6.United Nations. (1955). Principles for a Vital Statistics System. Statistical papers series, M(19). Check alsoWorld Health Organization. International Statistical Classification of Diseases and Related Health Problems,Tenth Revision (ICD-10), vol. 1. Geneva, 1992. (definition 3.1) (http://unstats.un.org/unsd/demographic/vital_statistics/definitions.htm)United Nations. (2002). A World Fit for Children: Resolution Adopted by the General Assembly. New York:United Nations. A/RES/S-27/2. Retrieved from http://www.unicef.org/specialsession/docs_new/documents/A-RES-S27-2E.pdfWasantwisut E. (2002). Recommendations for Monitoring and Evaluating Vitamin A Programs: OutcomeIndicators. Journal of Nutrition, 132:2940S-2942S.WHO. (1982). Control of Vitamin A Deficiency and Xerophthalmia. Geneva, Switzerland: World HealthOrganization. Report of a joint WHO/UNICEF/USAID/Hellen Keller International IVACG Meeting.WHO Technical Report Series 672:1-70.WHO. (1991). Indicators for Assessing Breastfeeding Practices. Geneva, Switzerland: World HealthOrganization. Report of an Informal Meeting. Division of Diarrhoeal and Acute Respiratory DiseaseControl, June 11-12.Growth Monitoring and Nutrition367

WHO. (1995). Physical Status: The Use and Interpretation of Anthropometry. Geneva, Switzerland: WorldHealth Organization. WHO Technical Report Series 854.WHO. (2001). Assessment of Iodine Deficiency Disorders and Monitoring Their Elimination: A Guide forProgramme Managers, 2nd Ed. Geneva, Switzerland: World Health Organization. WHO/NHD/01.1.WHO. (2001a). Iron Deficiency Anaemia: Assessment, Prevention and Control - A Guide for ProgrammeManagers. Geneva, Switzerland: World Health Organization. WHO/NHD/01.3.WHO. (2001b). New Data on the Prevention of Mother-to-Child Transmission of HIV and Their PolicyImplications: Conclusions and Recommendations. Technical Consultation on Behalf of the UNFPA/UNICEF/WHO/UNAIDS. Inter-Agency Task Team on Mother-to-Child Transmission of HIV. Geneva, Switzerland:World Health Organization. WHO/RHR/01.28.WHO. (2001c). Schistosomias and Soil-transmitted Helminth Infections. Geneva, Switzerland: World HealthOrganization. Fifty-fourth World Health Assembly, Agenda Item 13, WHA54.19. A54/VR/9. Retrievedfrom http://www.who.int/gb/EB_WHA/PDF/WHA54/ea54r19.pdfWHO. (2002). Report of the WHO Informal Consultation on the Use of Praziquantel and Albendazole/Mebendazole in Children Under 24 Months. Geneva, Switzerland: World Health Organization. WHO/CDS/CPE/PVC/2002.4.WHO/UNICEF. (1998). Complementary Feeding of Young Children in Developing Countries: A Review ofCurrent Scientific Knowledge. Geneva, Switzerland: World Health Organization. WHO/NUT/98.1.Wellstart International. Wellstart International's Tool Kit for Monitoring and Evaluating Breastfeeding Practicesand Programs.Yip R, and Ramakrishnan U. (2002). Experiences and Challenges in Developing Countries. Journal ofNutrition, 132:827S-830S.368Chapter 8

Annex 8.1. Complementary feeding questions from the DHS and KPC(1) A separate category for any foods made with red palm oil, palm nut, or palm nut pulp sauce must be added in countrieswhere these items are fed to young children. A separate category for any grubs, snails, insects or other small protein foodmust be added in countries where these items are fed to children. Items in each food group should be modified to includeonly those foods that are locally available and/or consumed in the country. Local terms should be used.(2) Grains include millet, sorghum, maize, rice, wheat, or other local grains. Start with local foods, e.g. ugali, nshima, fufu,chapati, then follow with bread, rice, noodles, etc.(3) Items in this category should be modified to include only vitamin A rich tubers, starches, or red, orange, or yellowvegetables that are consumed in the country.(4) These include cassava leaves, bean leaves, kale, spinach, pepper leaves, taro leaves, amaranth leaves, or other dark green,leafy vegetables.Chapter 8. Growth Monitoring and Nutrition369

9MORORTALALITYIndicators:• Late fetal death rate (cross-referenced in Chapter three)• Perinatal mortality rate (cross-referenced in Chapter three)• Cause-specific perinatal mortality rate (cross-referenced in Chapter three)• Birth weight specific mortality rate (cross-referenced in Chapter three)• Neonatal mortality rate (cross-referenced in Chapter three)• Infant mortality rate• Child mortality rate• Under-five mortality rate (cross-referenced in Chapter eight)• Cause-specific mortality rate

HAPTER 9. MCHAPTER9. MORORTALITALITYMortality reduction and improvements inchild health and nutrition are the ultimategoal of child health programs. Most of themortality impact is likely to occur throughreductions in deaths from five main causes:pneumonia, diarrhea, malaria, measles, andmalnutrition. As many child health programs donot include specific interventions targeted at themajor causes of death among children youngerthan seven days, the infant, child, and under-fivemortality rates have been the main mortalitymeasures of impact. However, many newbornhealth programs include interventions prior to thebirth of a child (during pregnancy, labor, anddelivery) and in the first few weeks of an infant’slife. Such interventions may include syphilisscreening; improving the diagnosis and treatmentof reproductive tract infections; expandingmaternal immunization with tetanus toxoid;including malaria prophylaxis during routineantenatal care; and educating community healthworkers to improve the identification andmalpresentation of labor and referral ofcomplicated cases to higher level facilities. In theearly weeks of life, interventions such as the useof resuscitation techniques for asphyxiated infants,proper management of neonatal sepsis and otherinfections; skin-to-skin kangaroo care for preterminfants, and the promotion of immediate andexclusive breastfeeding may also be introduced.Taking into account the diversity of child healthinterventions, this chapter addresses not only theconventional indicators of childhood mortality butalso the measures of fetal and early neonatalmortality that were cross-referenced in Chapterthree. We begin with a discussion of themethodological challenges of measuring fetal,infant, and child mortality.Measurement ChallengesLarge-scale survey data provide the most reliableestimates of child mortality, but they have limitations.Surveys such as the DHS/MICS/KPC arevaluable for tracking national infant and childmortality trends but they have a number oflimitations. First, such surveys are conducted onlyonce every three to five years (if that often).Second, in most countries, they do not yield preciseestimates for geographic sub-areas (such asdistricts or provinces), which is the level at whichprogram managers need the information. Third,these large-scale surveys are expensive to conductand analyze and cannot detect short-term changesin mortality rates.Few countries have sufficiently well developed vitalregistration systems that can yield valid and reliableinformation on all births and deaths in thecommunity. Health information systems can onlyprovide information on facility births and deathsand, in most settings, are poorly developed. Mostcommunity-based programs often do not have thecapacity to measure infant and child mortality.Concerning newborn health, prospective studieswould provide the most reliable mortality ratesbut are costly for regular reporting purposes.Large-scale surveys that rely on the retrospectivereporting of deaths in early infancy can provideestimates of neonatal mortality, but estimates ofperinatal mortality (which includes fetal deaths)require complete pregnancy histories. Manypopulation-based surveys in developing countriesfocus on obtaining birth histories.Chapter 9. Mortality371

The reliability of any survey estimate depends onthe completeness of reporting. Under-reportingtends to be more pronounced for deaths in earlyinfancy. Because of the relatively small numbersof deaths recorded in household surveys, nationalneonatal mortality rates are usually presented forthe five-year period before the survey, and subnationalestimates for the ten-year period beforethe survey. The lack of precision in the estimatessometimes makes it difficult to assess thesignificance of small changes between surveys(Rutstein, 1999).The quality of newborn and infant mortality data ispoorer than the quality of data for other ages.Evaluators require information on all births anddeaths to derive valid measures of newborn andinfant mortality. As mentioned previously, fewdeveloping countries have official vital registrationsystems that are reasonably reliable or complete.Institutional barriers leading to underreportingmay include cost, distance to the registration office,and a lack of awareness of the importance of birthand death registration. Social and cultural barriersmay also make it difficult to collect valid data.Even in countries with well-developed registrationsystems, biases towards reporting larger, olderbabies may exist, with deaths of small babies earlyin the neonatal period being omitted. Fetal deathsare much less likely to be reported than deaths oflive-born babies (WHO, 1996a).Definitions of newborn, infant, and child death areinconsistent.Many countries define and record births anddeaths in ways that may differ from the standarddefinitions recommended by WHO in ICD-10.Some countries, for example, only record a babyas a live birth if the baby survives beyond 24 hours(Lawn, McCarthy, and Ross, 2001). Thedefinition of a perinatal death is particularlyproblematic, not only because legal reportingrequirements may vary between countries, but alsobecause the two standard international definitionshave different gestational age and weight criteria(see the indicator Perinatal Mortality Rate).Other difficulties may arise because national birthand death criteria may be differentially interpretedand applied, leading to the misclassification of livebirths as fetal deaths and vice versa. This problemmay occur because of lack of training or experience(this is true of deaths that occur in facilities aswell as at home) or because of how institutionsand public-health authorities choose to interpretand apply national birth and death criteria.Changes in medical practice may also alter the waypractitioners systematically classify deaths.Ideally, all deaths (including fetal, perinatal, andneonatal) should be counted when estimating neonatalmortality, but in practice, counting neonatal deathsis often the only feasible approach.In any program, the types of deaths to be counteddepend on the program objectives, themeasurement method chosen, and the resourcesavailable. Ideally, program staff should recordinformation on all neonatal, perinatal, and fetaldeaths in order to derive a complete picture ofpregnancy outcomes. Having a complete pictureis important because the causes of stillbirths andneonatal deaths are often the same, and thedistinction between a stillbirth and a neonataldeath may, in practice, be a very fine one. Justexamining one rate or the other may underestimatethe true level of mortality in the newborn period.Counting all deaths will provide a measure of theimpact of interventions on etiologies that causeboth fetal and neonatal deaths and will reveal thechanges in the epidemiology of newborn deaths.Modern technologies, for example, may improvedelivery outcome but increase neonatal mortalitybecause fetal deaths are displaced from theantepartum to the postpartum period. Reportingneonatal, perinatal, and stillbirth rates together willalso permit more valid comparisons acrossprograms and settings. Realistically, however, fewprograms are able to achieve this goal as in manydeveloping countries, the majority of births anddeaths occur at home (WHO, 1996a).372 Chapter 9

Demonstrating the impact of child healthprograms on infant and child mortality requiresmore than simply tracking death rates over time.While many programs would like to attributeimprovements in child health and survival to theirprogram activities or interventions, in practice,many factors other than the program may havecontributed to these changes. Controlled fieldexperiments demonstrating what would havehappened in the absence of the program are oftennot feasible for evaluating ongoing national-levelprograms. Thus, an essential element of theevaluation of child health programs is to establishthe level of certainty that is needed for decisionmakers to act on the findings of the evaluation.Selection of IndicatorsThe indicators presented in this chapter are thefollowing:• Late fetal death rate• Perinatal mortality rate• Cause-specific perinatal mortality rate• Birth weight specific mortality rate• Neonatal mortality rate• Infant mortality rate• Child mortality rate• Under-five mortality rate• Cause-specific mortality rateNote that the first five indicators are cross-referencedin chapter three. Methods based on deaths avertedcan be also used to measure the impact of child healthprograms. These methods compare the observedperiod child mortality rates with hypothetical ratesthat would have been expected in the absence of aprogram. The difference between the observed andpotential mortality is used to derive an estimate ofthe number of deaths that did not occur during aspecified reference period due to the effect of aprogram.Mortality373

LATEFETETALDEAEATHRATETETE (LFDR)DefinitionNumber of late fetal deaths per 1,000 births (livebirths plus late fetal deaths) in a given period.Numerator: Number of babies born dead after28 weeks of gestation (or birth weight over 1kg) in a given period.Denominator: Total number of births (livebirths plus fetal deaths) in the same period.A “late fetal death” is defined as death of a fetusafter 28 weeks of gestation.The WHO definition of a “fetal death,” alsoadopted by the United Nations and the NationalCenter for Health Statistics (NCHS), is deathbefore the complete expulsion or extraction fromits mother of a product of conception, irrespectiveof the duration of pregnancy. The death isindicated by the fact that after such separation,the fetus does not breathe or show any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitivemovement of voluntary muscles.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extraction fromits mother of a product of conception, irrespective ofthe duration of pregnancy, which, after suchseparation, breathes or shows any other evidence oflife, such as beating of the heart, pulsation of theumbilical cord, or definite movement of voluntarymuscles, whether or not the umbilical cord has beencut or the placenta is attached; each product of sucha birth is considered live born.”The terms “stillbirth” and “fetal death” aresometimes used interchangeably.Measurement ToolsPopulation-based surveys; vital registration; servicestatistics; routine HISWhat It MeasuresThe LFDR directly reflects prenatal andintrapartum care. It is a measure of the quality ofmaternal health care services. There are anestimated four million late fetal deaths each year.Common causes of late fetal death are preventablematernal STIs such as syphilis, intrapartum birthasphyxia, pre-term birth, birth defects, maternalhypertension, and maternal diabetes.How to Measure ItThis indicator requires two pieces of informationfor a given population and reference period: thenumber of live-born babies and the number of fetaldeaths from 28 weeks gestation (equivalent tobirth weight over 1 kg). WHO usually refers tothe expected weight at a given gestational age sincegestational assessment is often unavailable (WHO,1992). At the facility level, the numerator can bemeasured from birth registers or delivery room logsand from case reviews at the health facility (or inthe community). Some countries, such asMalaysia, have nationwide systems for reportinglate fetal deaths.The denominator is the number of live births plusthe number of fetal deaths occurring in the samereference period. Where data on the number oflive births are unavailable, evaluators can estimatethe total number of live births using census datafor the total population and crude birth rates in aspecified area as follows:Total expected births= population x crude birth rateMortality375

In settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (i.e., number of live births orpregnant women = total population x 0.035[WHO, 1999a, 1999b]).Fetal deaths or stillbirths are under-recorded inmany settings. However, they can be estimatedby using the ratio of stillbirths to early neonataldeaths, which is usually 1:1. Thus, if the numberof deaths during the first week of life is known,the number of late fetal deaths can be estimatedas this will be equal to the number of neonataldeaths. In some situations, such as high rates ofsyphilis infection, there may be more stillbirthsthan neonatal deaths.Strengths and LimitationsThis indicator suffers from underreporting andunder-recording. Health facility-based datasignificantly underestimate the problem of latefetal deaths because in many settings, many latefetal deaths and live births occur outside the healthsystem, which will cause substantial selection bias.Therefore, facility data are not recommended forestimating the LFDR for the general population.Pregnancy histories, now included in manysurveys, including the DHS, are another sourceof data for calculating this indicator. However,there has been relatively less experience withpregnancy histories than with birth historiesbecause of concerns about the quality ofretrospectively reported pregnancy histories.Common problems with data quality include theomission of late fetal and early neonatal deathsand difficulty in obtaining accurate informationon gestational age or birth weight leading to themisclassification of some stillbirths as latespontaneous abortions.International comparisons are limited bynonstandard definitions and terminology for latefetal deaths and highly variable application of thesedefinitions. For example, ICD-10 defines late fetaldeaths from 22 to 40-weeks gestation. Forinternational comparisons, a birth weight of atleast 1000g (or of 28-weeks gestation or more ifweight is unavailable) is recommended. Errors orconfusion may arise in distinguishing between alive birth and a late fetal death.Both the coverage and quality of data on late fetaldeaths are insufficient and unreliable in manydeveloping countries due to sociocultural reasons,health system barriers, and the poor coverage andquality of vital registration systems. Socioculturalbarriers to obtaining information on pregnancyand birth outcomes include seclusion of womenand newborns at home, misconceptions about theimportance of registration and data collection, andacceptance of fetal-neonatal deaths as normal.Barriers within the health system include the lackof motivation for staff to collect necessary data,perceptions of the viability of the baby, and errorsin the coding of cause of death. Barriers toregistration include issues of accessibility andaffordability and lack of awareness of the benefitsof registration (Lawn et al., 2001).376Chapter 9

PERINERINATALAL MORORTALITALITY RATETETE (PMR)DefinitionNumber of perinatal deaths per 1,000 total births(live births and fetal deaths) in a given period.Numerator: Sum of fetal deaths and deaths tolive-born babies within the first sevencompleted days (i.e., age 0-6 days) of life in agiven period.Denominator: Total number of births (livebirths and fetal deaths) in the same period.A “perinatal death” is a fetal death or an earlyneonatal death.The WHO definition of a “fetal death,” alsoadopted by the United Nations and the NationalCenter for Health Statistics (NCHS), is deathbefore the complete expulsion or extraction fromits mother of a product of conception, irrespectiveof the duration of pregnancy. The death isindicated by the fact that after such separation,the fetus does not breathe or show any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitivemovement of voluntary muscles.A “fetal death” is the death of a fetus weighing 500gor more, or of 22-weeks gestation or more if weightis unavailable (ICD-10). The terms stillbirth andfetal death are sometimes used interchangeably.A “late fetal death” is defined as death of a fetusafter 28 weeks of gestation.An “early neonatal death” (END) is the death of alive newborn within the first seven completed days(i.e., 0-6 days) of life. Note: The day of birth iscounted as day 0, so that “within the first 7completed days” or “within 1 week” includes babies0-6 days old.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Great variation exists both between and withincountries on how the fetal death component ofperinatal mortality is recorded, particularly forearly fetal deaths that occur at 22 to 27-weeksgestation. For international comparison, WHO(1992) suggests including only deaths of fetusesweighing at least 1,000g or of 28-weeks gestationor more if weight is unavailable. Presentations ofthe PMR should include a clear statement of thedefinition of perinatal mortality used. In practice,in most developing countries, accurate data onbirth weight or gestational age are difficult toobtain.Measurement ToolsPopulation-based surveys; vital registration; servicestatistics; routine HISWhat It MeasuresThe PMR is a key outcome indicator for newborncare and directly reflects prenatal, intrapartum, andnewborn care. It is estimated that perinatal deathsaccount for approximately 7% of the global burdenof disease (World Bank, 1993). The early neonatalcomponent of the PMR may respond relativelyquickly to programmatic interventions, forexample, following the introduction of elementsof the WHO “Essential Newborn Care Package.”The fetal death component may decline moreMortality377

slowly because it depends more on interventionsthat primarily influence maternal health and onthe availability of technologies such as caesariansection.How to Measure ItThis indicator requires three pieces of information:the number of fetal deaths in a given populationin a given period (i.e. 12 months); the number ofdeaths of live-born babies at age 0-6 days in thesame population and period; and the number ofbirths (live births plus fetal deaths) in the samepopulation and period.The PMR obtained from large population-basedsurveys may be calculated at the subnational levelif sample sizes are sufficiently large.Strengths and LimitationsBecause the PMR includes both fetal deaths anddeaths in the first week of life, it avoids theproblems of defining a live birth. There are,however, problems with the identification of a fetaldeath or stillbirth. According to the WHO, a fetaldeath occurs after the twenty-second week ofgestation. However, different countries often useslightly different definitions, making internationalcomparisons of the PMR difficult.In many countries, vital registration data are notsufficiently complete to allow reliable estimationof the PMR. Techniques now exist for collectingdata on stillbirths, live births, and early neonataldeaths in population-based surveys throughpregnancy histories. These pregnancy historiesare now included in many surveys, including theDHS. However, there has been relatively lessexperience with pregnancy histories than withbirth histories because of concerns about thequality of retrospectively reported pregnancyhistories.Data quality is an issue. Common problems withdata quality include:378• Omission of stillbirths and early neonataldeaths. It is estimated that perinatal mortalityrates are under-reported by at least 40%(WHO and UNICEF, 1996) and by as muchas 500% in countries with high death rates(Lumbignon, Panamonta, Laopaiboon,Pothinam, and Patithat, 1990; McCaw-Binns,Fox, Foster-Williams, Ashley, and Irons,1996);• Difficulty in obtaining accurate informationon gestational age or birth weight leading tothe misclassification of stillbirths as latespontaneous abortions; and• Heaping of the reported age at death of livebirths on seven days, leading to the misclassificationof early neonatal deaths aslate neonatal deaths.*Prospective population-based surveys of pregnantwomen provide better quality data, but areexpensive to undertake.Survey-based estimates are generally subject torelatively large sampling errors making it difficultto detect changes over short periods unless thechanges are quite large. Retrospective surveybasedestimates are often based on a five-yearperiod prior to the survey.The following caveats bear mention. The PMRis sensitive to changes in the quality of data. Forexample, a rise in the PMR may indicatedeterioration in perinatal outcomes, or animprovement in the reporting of perinatal deaths.Therefore, an assessment of data quality is anessential component of analysis. In this context,evaluators often find it useful to disaggregate thePMR into its two components: stillbirths and* Heaping occurs when respondents do not know the exactage at death. Estimated ages at death are often reported oncertain preferred ages, such as seven days, leading to adistorted age distribution of deaths in which too manydeaths are reported at the preferred age, and too few at theages just before and after.Chapter 9

early neonatal deaths. Data quality is generallymore problematic for stillbirths than for earlyneonatal deaths, because of the ambiguity over thedefinition of fetal deaths and problems obtaininggestational age (WHO, 1996a).Facility data are not recommended for estimatingthe PMR for the general population because inmany settings, many perinatal deaths and livebirths occur outside the health system. Facilitybasedestimates of the PMR should also beinterpreted with caution because the rate issensitive to the types of deliveries occurring in thefacility. Consequently, the PMR may rise or fallin response to changes in the mix of deliveries in afacility. In small facilities, the PMR will bepotentially unstable because of the small numberof deliveries and perinatal deaths; thus, the PMRmay be ineffective for monitoring change over timein a single facility. Facility-based data are moreuseful for monitoring in countries where a largeproportion of births take place in facilities andwhere the completeness of routine reporting ishigh.Mortality379

-SCAUSEUSE-S-SPECIFICPERINERINATALAL MORORTALITALITY RATEDefinitionNumber of perinatal deaths from a specific causeper 1,000 births (live births and fetal deaths) in agiven period.Numerator: Number of perinatal deaths froma specific cause in a given period x 1,000(or 10,000 or 100,000).Denominator: Total number of births (livebirths and fetal deaths) in the same period.A “perinatal death” is a fetal death or an earlyneonatal death.The terms “stillbirth” and “fetal death” aresometimes used interchangeably.Specific causes that are commonly measuredinclude:• Lethal or severe congenital abnormalities;• Acute intrapartum events, resulting inintrapartum stillbirths or neonatal deaths dueto “asphyxia;”; and• Infections, which may be highly specific, e.g.,syphilis infections as a cause of stillbirthsand early neonatal deaths.The exact causes of death being coded may differdepending upon the program or locality where theindicator is being collected.The WHO definition of a “fetal death,” alsoadopted by the United Nations and the NationalCenter for Health Statistics (NCHS), is deathbefore the complete expulsion or extraction fromits mother of a product of conception, irrespectiveof the duration of pregnancy. The death isindicated by the fact that after such separation,the fetus does not breathe or show any otherMortalityevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitivemovement of voluntary muscles.A “fetal death” is the death of a fetus weighing 500gor more or of 22-weeks gestation or more if weightis unavailable (ICD-10). The terms stillbirth andfetal death are sometimes used interchangeably.An “early neonatal death” (END) is the death of alive newborn within the first seven completed days(i.e., 0-6 days) of life. Note: The day of birth iscounted as Day 0, so that “within the first 7completed days” or “within 1 week” includes babies0-6 days old.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Great variation exists both between and withincountries on how the fetal death component ofperinatal mortality is recorded, particularly forearly fetal deaths that occur at 22 to 27-weeksgestation. For international comparison, WHO(1992) suggests including only deaths of fetusesweighing at least 1000g or of 28-weeks gestationor more if weight is unavailable. Presentations ofthe PMR should include a clear statement of thedefinition of perinatal mortality used. In practice,in most developing countries, accurate data onbirth weight or gestational age are difficult toobtain.381

Measurement ToolsFacility-based perinatal death audits; community/demographic surveillance; vital registration; verbalautopsyWhat It MeasuresThis indicator measures death from specific causesduring the perinatal period. Measurement ofcause-specific perinatal mortality is important forseveral reasons, including the following: (1) toestablish the relative public health importance ofthe different causes of death; (2) to evaluate trendsover time for specific causes of death; (3) toevaluate health interventions aimed at reducingmortality from specific causes of death; (4) toinvestigate the circumstances surrounding thedeaths in order to identify ways to reduceunnecessary death; and (5) to facilitate researchinto factors associated with mortality from specificcauses of death.How to Measure ItThis indicator can be measured at the healthfacility, district, national or community levels,(although death coding at the community levelmay be less reliable), or a more simple surrogatemeasure of a specific cause may need to be used.Perinatal Death Audits (PNDAs) are beingpromoted increasingly in developing countries,particularly at the facility level. Several trainingprograms are available for PNDAs withsupporting software, including the PerinatalEducation Program, a distance education programavailable on the internet that has been used to trainover 30,000 doctors and nurses, largely in SouthAfrica (www.pepcourse.co.za), with the PerinatalProblem Identification Program Software used toanalyze the data (www.ppip.co.za).In areas where medical certification of cause ofdeath is rare, verbal autopsy is often used toidentify the causes of death among infants andchildren. Demographic surveillance, where alldeaths are reported on a regular basis throughoutthe year, may be used to identify deaths.In verbal autopsy, differing methods can be usedto obtain a verbal account of the cause of death.In an open-ended history, the caregiver or nextof-kinis asked to tell about the events leading upto the child’s death in their own words and probedto follow-up on particular aspects. Close-endedquestions ask whether specific symptoms and signswere present during the final illness. “Expert”opinions or computerized algorithms are thenapplied to allocate the presumed cause of deathusing the descriptive data.Strengths and LimitationsIt is difficult to assign causes of death, even at thehealth facility level. At the community level, datacollection is often retrospective and reliant uponverbal autopsy, rather than on a clinicallydetermined cause of death. These factorscontribute to bias and may make the validity andreliability of the data questionable. Failure toenumerate all deaths can lead to an invalid measureof proportionate cause of death. For example,selective undercounting of deaths in the first hourof life will disproportionately reduce the numbersof deaths due to asphyxia and severe preterm birth.Unless the neonatal mortality rate is high, and/ora large number of births are included, the causespecificmortality rate may be misleading, as smallnumbers will not allow for the investigation of timetrends.Vital registration systems often do not havesufficient coverage to provide accurate data aboutcause-specific mortality in developing countries.Usually a 90% coverage rate is taken as a cut–offfor representation. Demographic surveillancetends to cover limited geographic areas, thus theunderlying cause-specific mortality in these areascannot be necessarily generalized to widerpopulations, as some of the populations undersurveillance are not typical due to multiple trialsand interventions.Misclassification of the cause of death not onlyaffects estimates of levels of cause-specificmortality over time, but also comparisons in cause-382Chapter 9

specific mortality rates between two populationgroups. In mortality surveys, the accuracy of theindicator depends on the ability of respondents todescribe the final illness as well as on the way inwhich diseases are understood and described inthe community. Clear case definitions and theuse of hierarchical categories for allocating causeof death will minimize subsequent errors.One limitation of cause-specific mortality rates isthat the death of a child is commonly the result ofmore than one cause. Some verbal autopsyquestionnaires, such as those developed by WHO,Johns Hopkins School of Hygiene and PublicHealth, and the London School of Hygiene andTropical Medicine, allow for multiple causes ofdeath, while others only allow for one. Wheninterpreting this indicator, it is important to knowwhether multiple causes of death are allowed forin the coding since the sum of the proportions foreach cause of death will generally be greater than1.00 when multiple causes of death are allowed.For this reason, many analyses do restrict the majorcause of death to one cause per child. The PerinatalProblem Identification Program in South Africaallows for the coding of primary obstetric causesof death (for stillbirths and neonatal deaths), finalcauses of death (in neonatal deaths), and avoidablecauses of death that are patient related, health careworker related, and administrative.Mortality383

BIRIRTHWEIGHTSPECIFICMORORTALITALITY RATE(BWSMR)DefinitionThe Birth Weight Specific Mortality Rate(BWSMR) is a stratification of a “newbornmortality rate” by birth weight grouping. Forexample, the Birth Weight Specific NeonatalMortality Rate (BWSNMR) for births weighing2500g or more is calculated as:Numerator: Number of neonatal deathsweighing 2500g or more at birth.Denominator: Total number of live birthsweighing 2500g or more at birth.And for births under 2500g, the BWSNMR iscalculated as:Numerator: Number of neonatal deathsweighing under 2500g at birth.Denominator: Total number of live birthsweighing under 2500g at birth.Measurement ToolsService statistics; HIS (in highly developedsystems)What It MeasuresBirth weight is the most sensitive predictor ofinfant survival and a good predictor of maternalhealth and well-being. The mortality rate for lowbirth weight babies is much higher than for thosewith a normal birth weight. Stratifying newborndeaths by birth weight helps to determine thecause of death and therefore to identify whereinterventions are needed. For example, deaths ofvery small babies are more likely related tomaternal causes predisposing to intrauterinegrowth retardation and preterm birth, whereasdeaths of normal birth weight babies are morelikely to be related to intrapartum asphyxia andpoor obstetric care. In the first case, interventionsshould focus on the mother (improving nutritionand reducing antenatal infection) and, in thesecond case, the focus should be on improving thequality of delivery care. Evaluators can obtainadditional information by stratifying birth weightby time of death (see Table 3.5 on page 102).How to Measure ItThe data requirements for calculating thisindicator are the number of deaths in a particularbirth weight grouping and the total number ofbirths in the same weight grouping.Strengths and LimitationsInformation for this indicator can only be collectedin settings where all babies are weighed. It istherefore most appropriate for use in healthfacilities but has been collected in somecommunity settings as part of maternal andperinatal health area surveillance systems (Lawnet al., 2001).One useful application of this type ofdisaggregation is to examine the number ofintrapartum deaths in normal birth weight babies.If the quality of obstetric care is good (and womenare not presenting very late in labor), then veryfew intrapartum deaths should occur becausedeliveries are expedited rapidly. The proportionof fetal deaths in babies of normal birth weightmay serve as a proxy indicator for intrapartumasphyxia and quality of delivery care.Mortality385

NEONEONATALAL MORORTALITALITY RATETETE (NMR)DefinitionNumber of neonatal deaths per 1,000 live birthsin a given period.Numerator: Number of deaths within the first28 completed days of life (0-27 days) in a givenperiod x 1000.Denominator: Total number of live births inthe same period.The NMR is often broken down into early andlate neonatal mortality rates. The early neonatalmortality rate (ENMR) is calculated as follows:Numerator: Number of deaths within the firstseven completed days of life (0-6 days) in agiven period x 1000.Denominator: Total number of live births inthe same period.The late neonatal mortality rate (LNMR) iscalculated as follows:Numerator: Number of deaths within 7-27completed days in a given period x 1000.Denominator: Total number of live births inthe same period.A “neonatal death” is defined as a death within thefirst 28 completed days of life (0-27 days).A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Note: The day of birth is counted as day 0, so that“within the first 7 completed days” or “within 1week” includes babies 0-6 days old.Measurement ToolsCensus; population-based surveys (e.g., DHS,MICS, KPC); vital registration system; servicestatisticsWhat It MeasuresThe NMR is a key outcome indicator for newborncare and directly reflects prenatal, intrapartum, andneonatal care. Early neonatal deaths are moreclosely associated with pregnancy-related factorsand maternal health, whereas late neonatal deathsare associated more with factors in the newborn’senvironment.How to Measure ItTo calculate this indicator, two pieces ofinformation are needed: the number of neonataldeaths in a given population and reference period,and the number of live births in the samepopulation and reference period. The referenceperiod is usually one year but it could also be fiveyears.Where data on the numbers of live births for thedenominator are unavailable, evaluators cancalculate total estimated live births using censusdata for the total population and crude birth ratesin a specified area:Total expected births = population x crude birth rateMortality387

In settings where the crude birth rate is unknown,WHO recommends using 3.5% of the totalpopulation as an estimate of the number ofpregnant women (number of live births orpregnant women = total population x 0.035[WHO 1999a; WHO 1999b]).Routine HIS may collect data for this indicatorto obtain estimates of the NMR for facilities.Facility data are not recommended for estimatingthe NMR for the general population, because inmany settings, a number of neonatal deaths andlive births occur outside the health system, whichwill cause substantial selection bias.The NMR is usually calculated at the nationallevel. Sub-national estimates can also be calculatedif sample sizes are sufficiently large. The NMR issometimes calculated at a facility level to monitorthe outcome of delivery and newborn care inhealth facilities. Reliable estimates for individualfacilities can only be obtained for very largefacilities if there are large numbers of deliveriesand neonatal admissions.Strengths and LimitationsIn many countries, vital registration data are notsufficiently complete to allow reliable estimation ofthe NMR. The standard techniques for collectingdata on live births and neonatal deaths in populationbasedsurveys have been widely applied in programssuch as the WFS and DHS. Data quality is animportant issue; common problems include omissionof deaths, particularly very early neonatal deaths, andheaping of the reported age at death on seven, 28, or30 days.* Heaping on these digits is particularlyproblematic because it will lead to themisclassification of early neonatal deaths as lateneonatal death (seven days) or late neonatal deathsas post-neonatal deaths (28 and 30 days).The NMR may respond fairly quickly toprogrammatic interventions, for example,immunizing all pregnant women in areas of hightetanus prevalence. However, survey-basedestimates are generally subject to relatively largesampling errors, so it is impossible to detectchanges over short periods of time unless thechanges are quite large. Also, changes in neonatalmortality rates are usually a long-termphenomenon and thus occur slowly. Therefore,we recommend collecting estimates of the NMRevery three to five years or longer.One limitation is that the NMR is sensitive tochanges in data quality. For example, a rise in theNMR may indicate deterioration in newbornhealth outcomes, or it may indicate animprovement in the reporting of neonatal deaths.Therefore, assessing data quality is essential toanalysis.Also, comparisons of facility-based estimates ofthe NMR should be interpreted carefully becausethe NMR in a facility is very sensitive to the casemix of deliveries and neonatal admissions. Ahigher NMR in one facility may not suggestpoorer quality of neonatal care in that facilitybecause the NMR may rise or fall with changes inthe case-mix. Also, improvements in prenatal andintrapartum care and advances in medicaltechnology may increase the NMR because babieswho may otherwise have been stillbirths maysurvive delivery only to die in the neonatal period.For these reasons, we recommend that evaluatorsbreak down facility-based estimates of the NMRby birth weight (see Birth Weight SpecificMortality Rate) and by admission status (directadmission or transfer-in) as a proxy for case mix.* Heaping occurs when respondents do not know the exactage at death. Estimated ages at death are often reported oncertain preferred ages, such as seven, 28, or 30 days, leadingto a distorted age distribution of deaths in which too manydeaths are reported at these preferred ages, and too few atthe ages just before and after.388Chapter 9

INFNFANTMORORTALITALITY RATETETE (IMR)DefinitionThe number of deaths to infants (children underone year of age) per 1,000 live births in a givenperiod.Numerator: Number of deaths to infants(children under one year of age) in a givenperiod x 1,000.Denominator: Total number of live births inthe same period.The infant mortality rate can also be expressed asa probability of a child born in a specified yeardying before reaching the age of one year.“Infancy” is regarded as the first year of life.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Measurement ToolsCensus; population-based surveys (e.g., DHS,MICS, KPC); vital registration systemWhat It MeasuresThe infant mortality rate measures death in thefirst year of life. It captures both deaths that are aresult of genetic and structural malformations andbirth delivery complications (most frequently inthe first month of life) and those that are associatedmore with external social or environmentalconditions (in the remaining months of the firstyear of life)How to Measure ItThe infant mortality rate is calculated by dividingthe number of infant deaths (i.e., the deaths ofchildren under one year of age) in a given periodby the total number of live births in the sameperiod and multiplying the result by 1000. Aninfant death can be recorded only when a baby isborn alive and when a baby born alive dies beforereaching its first birthday. If there has been a stillbirth,then it would be classified as a fetal deathand not be included in either official birth statisticsor death statistics.The infant mortality rate can be calculated from avital registration system – one covering at least90% of vital events in the population. This levelof completeness is not found in national vitalregistration systems in many developing countries.Delays in processing data and difficulties inestimating the appropriate denominators tocalculate the rate may also limit the use of vitalregistration data for measuring this indicator.In many countries, sample surveys such as theDHS and censuses are the primary sources ofinformation for the calculation of the infantmortality rate. Censuses and surveys usually askquestions on the number of children ever born,the survivors by age, and the age of the mother.From these data, one can estimate infant mortalityrates indirectly using the Brass Method, as in theMICS (see United Nations Manual X [UN, 1983]for a complete description of this technique).These estimates are generated in the form of atime series trend and not as single estimates forindividual points in time. Indirect estimates areless appropriate for measuring program impact,Mortality389

since the entire time series trend is reevaluatedwhen new empirical data become available.In the DHS, infant mortality is calculated directlyfrom the birth histories of women. Women areasked questions about each live birth they haveever had. These questions cover the date of birth,current age, survival status of the child, and, if thechild is dead, the age at death. From thisinformation it is possible to construct life tableestimates of mortality.The Programme of Action of the InternationalConference on Population and Developmentencouraged countries with intermediate mortalitylevels to achieve an infant mortality rate below 50deaths per 1,000 births by the year 2005, and allcountries to achieve an infant mortality rate below35 per 1,000 live births by 2015. The IMR can becalculated every year from a vital registrationsystem – one covering at least 90% of vital eventsin the population. When data are derived fromhousehold surveys, the IMR should be measuredevery five years.Strengths and LimitationsChanges in mortality rates are usually a long-termphenomenon and thus occur slowly. Therefore,for purposes of monitoring change, this indicatorshould be measured at least every five years. Theindicator is not recommended for use at the locallevel as it may lead to unrealistic programexpectations. When interpreting trends inmortality, it is important not to combine indirectestimates with direct estimates in order to formtrends lines because of the possibility of spurioustrends, which may be an artifact of themethodologies used to estimate the rates anddifferences in the reporting periods.Sometimes, it is difficult to determine whether afetus has been born live or dead. Thisdetermination is often based on a judgment by anattending physician, nurse, or midwife and has aninfluence of infant mortality measures. Ambiguityabout whether a live birth has occurred hasconsequences for whether the death is classifiedas a fetal death or death to a live-born.In developing countries, household surveys areessential to the calculation of this indicator, butthere are some limits to their quality. Survey dataon under-five deaths are subject to recall error, andrequire large samples, which may be costly.Therefore, when using household surveys it isimportant to take into account sampling errors.In addition, indirect estimates rely on estimatedactuarial (“life”) tables that may be inappropriatefor the population concerned.The indirect technique for calculating infantmortality requires fewer and simpler questionsresulting in substantial cost savings. It also requiresthe collection of birth histories, which may bedifficult and time consuming. However, birthhistories give the most robust estimates of infantmortality, short of actual birth and deathregistration. Vital registration data may sufferfrom a number of problems, including omissionof some births and deaths, especially infants thatdied shortly after birth, and the misreporting ofthe date of birth and the age at death. Omissionof infant deaths is usually most severe for deathsthat occur early in infancy.Clearly, risks of death to infants are greatest atthe instant after birth and may declinesubsequently. For this reason, it is useful to divideinfant mortality into components, according to thetime of death. The most frequently useddistinctions are deaths in the first month ortwenty-eight completed days of life (calledneonatal mortality) and deaths in the remainingmonths of the first year of life (called postneonatalmortality). When this distinction is used, itrecognizes the different patterns of causes of deathin each period. While neonatal deaths frequentlyarise out of conditions originating in the prenatalperiod, postneonatal deaths are more likely toreflect the socioeconomic conditions of the home(e.g., quality of care and nutrition), as well asinfectious diseases and other causes. In calculating390Chapter 9

the postneonatal mortality rates, the numerator isthe number of deaths of children from 28 days upto, but not including, one year of age during a giventime period and the denominator is the numberof live births during the same period. Thepostneonatal mortality rate is usually expressed per1,000 live births.It is to be noted that in settings with high AIDSprevalence, estimates of infant and child mortalityusing the direct or indirect approach may becompromised. This bias is due to the fact thatchildren whose mothers have died tend to havehigher mortality, and in surveys, data on childdeaths are collected retrospectively from livingmothers.Mortality391

CHILDMORORTALITALITY RATETETE (CMR)DefinitionThe number of deaths to children aged 12-59months per 1,000 children in this age group in agiven period.Numerator: Number of deaths to children aged12-59 months in a given period x 1,000.Denominator: Total number of children aged12-59 months in the same period.The CMR may also be expressed as the probability(expressed as a rate per 1,000 live births) of a childborn in a specified year dying between age oneyear and the fifth birthday if subject to currentage-specific mortality rates. When the CMR ispresented as the probability of dying between theexact age of one year and the fifth birthday, it isexpressed as 4q 1.Measurement ToolsCensus; population-based surveys (e.g. DHS);vital registration systemWhat It MeasuresThe child mortality rate measures the risk of deathin the first several years of life. It captures deathsthat are associated more with external, social, orenvironmental conditions, such as nutrition,sanitation, communicable diseases of childhood,and accidents occurring in and around the home.The child mortality rate also reflects the level ofpoverty and is consequently a sensitive indicatorof the level of socioeconomic development. Whilethis indicator is less responsive to rapid changesin the health status of infants and children, it isindicative of the broader range of socioeconomicconditions that affect overall health.How to Measure ItChild mortality is measured by relating deaths tochildren aged 12-59 months (in the numerator)to the total population of children aged 12-59months and then multiplying by a constant, usually1,000. CMR is usually calculated for a specificperiod of time. The period may be as short as oneyear, though a longer period of three to five yearsis typically used. In most surveys, including DHS,estimates are often for five-year periods.Child mortality data can be derived directly fromvital registration systems – one covering at least90% of vital events in the population. These areoften not available in developing countries andmay not be a valid source of information inlocations where the coverage of vital registrationsystems is low. Delays in processing data anddifficulties in estimating the appropriatedenominators to calculate the rate may also limitthe use of vital registration data for measuring thisindicator.In many countries, sample surveys such as theDHS and censuses are usually the primary sourcesof information for the calculation of the infantmortality rate. Censuses and surveys usually askquestions on the number of children ever born,the survivors by age, and the age of the mother.From these data, one can estimate child mortalityrates indirectly using mathematical modeling.These estimates are generated in the form of atime series trend and not as single estimates forindividual points in time (see United NationsManual X [UN, 1983] for a complete descriptionof this technique). Indirect estimates are lessappropriate for measuring program impact, sincethe entire time series trend is reevaluated whennew empirical data become available.Mortality393

In the DHS, child mortality is calculated directlyfrom the birth histories of women. Women areasked questions about each live birth they haveever had. These question cover the date of birth,current age, survival status of the child, and, if dead,age at death. From this information, it is possibleto use life table analysis to construct directestimates of child mortality.Goal 4, Target 5 of the Millennium DevelopmentGoals (MDG) is to reduce by two-thirds themortality rate among children under five by 2015.The CMR can be measured annually where vitalregistration systems cover at least 90% of vitalevents in the population. When data are derivedfrom household surveys, the CMR should bemeasured every five years.Strengths and LimitationsChanges in mortality rates are usually a long-termphenomenon and thus occur slowly. Therefore,for purposes of monitoring change, this indicatorshould be measured at least every five years. Theindicator is not recommended for use at the locallevel as it may lead to unrealistic programexpectations. When interpreting trends inmortality, it is important not to combine indirectestimates with direct estimates in order to formtrends lines because of the possibility of spurioustrends, which may be an artifact of themethodologies used to estimate the rates anddifferences in the reporting periods.In developing countries, household surveys areessential to the calculation of this indicator, butthere are some limits to their quality. Survey dataare subject to recall error, and surveys estimatingunder-five deaths require large samples, whichmay be costly. Therefore, when using householdsurveys it is important to consider sampling errors.In addition, indirect estimates rely on estimatedactuarial (“life”) tables that may be inappropriatefor the population concerned.In settings with high AIDS prevalence, estimatesof child mortality using the direct or indirectapproach may be compromised. This bias is dueto the fact that children whose mothers have diedtend to have higher child mortality, and in surveys,data on child deaths are collected retrospectivelyfrom living mothers.The indirect technique for calculating childmortality requires fewer and simpler questions,resulting in substantial cost savings. Collectingbirth histories may be difficult and timeconsuming. However, information from birthhistories provide the most robust estimates ofinfant mortality, short of actual birth and deathregistration. Vital registration data may sufferfrom a number of problems, including omissionof some births and deaths, especially infants thatdied shortly after birth, and the misreporting ofthe date of birth and age at death. Omission ofinfant deaths is usually most severe for deaths thatoccur early in infancy.394Chapter 9

-FUNDERNDER-F-FIVEMORORTALITALITY RATETETE (U5MR)MDG G IndicatorDefinitionThe number of deaths to children aged 0-4 yearsin a given period per 1,000 live births in the sameperiod.Numerator: Number of deaths to children aged0-4 years in a given period x 1,000.Denominator: Total number of live births inthe same period.The U5MR may also be expressed as theprobability (expressed as a rate per 1,000 livebirths) of a child born in a specified year dyingbefore reaching the age of five if subject to currentage-specific mortality rates. When under-fivemortality is presented as a probability of dyingbefore age five, it is expressed as 5q 0. This is thedefinition typically used when calculating U5MRfrom household surveys. Strictly speaking this isnot a rate (i.e., the number of deaths divided bythe total population at risk during a given periodof time), but a probability of death derived from amodel life table and expressed as the rate per 1,000live births.A “live birth” is described by the United Nations(2001) as “the complete expulsion or extractionfrom its mother of a product of conception,irrespective of the duration of pregnancy, which,after such separation, breathes or shows any otherevidence of life, such as beating of the heart,pulsation of the umbilical cord, or definitemovement of voluntary muscles, whether or notthe umbilical cord has been cut or the placenta isattached; each product of such a birth is consideredlive born.”Measurement ToolsCensus; population-based surveys (e.g., DHS,MICS); vital registration systemWhat It MeasuresThis indicator is a composite of the risks of dyingin infancy and early childhood. It also reflects thesocial, economic, and environmental conditions inwhich children (and others in society) live,including their health care. The under-fivemortality rate captures more than 90% of globalmortality among children under the age of 18. Theindicator may be used, therefore, as a measure ofchildren’s well-being and the level of effort beingmade to maintain child health.How to Measure ItAt the national level, the best source of data is acomplete vital statistics registration system – onecovering at least 90% of vital events in thepopulation. Such systems are uncommon indeveloping countries, so estimates are alsoobtained from sample surveys or derived byapplying direct and indirect estimation techniquesto vital registration, census, or survey data.A wide variety of household surveys, includingthe MICS and DHS, are used in developingcountries to estimate under-five mortality rates.In the DHS, birth histories are collected fromwomen aged 15-49 years. Each eligible woman isasked the dates of birth and, where relevant, theage of death of every live-born child she has had.Using life-table analysis, this approach producesdirect estimates of under-five mortality.A second survey method is called the Brassmethod, which asks each woman surveyed simplequestions such as her age, the total number ofchildren she has borne, and the number of thoseMortality395

children who have died. The age of the motherand the proportion of children that died, inconjunction with assumption of fertility and otherfactors, allow an estimation of under-five mortality.This approach is used by the MICS and producesindirect estimates of under-five mortality (for acomplete description of this technique, see UnitedNations Manual X [UN, 1983]). Longitudinal orprospective surveys following a sample of babiesborn in a specified year over a number of yearsmay also be used to estimate under-five mortality.However, this approach is relatively uncommonin developing countries.Vital statistics are typically available once a year,but they are unreliable in most developingcountries. Household surveys that includequestions on births and deaths are generallyconducted every three to five years. Target 5 ofthe Millennium Development Goals (MDG) isto reduce the under-five mortality rate by twothirdsbetween 1990 and 2015. The U5MR canbe measured annually where vital registrationsystems cover at least 90% of vital events in thepopulation. When data are derived fromhousehold surveys, the U5MR should be measuredevery five years.Strengths and LimitationsChanges in mortality rates are usually a long-termphenomenon and thus occur slowly. Therefore,for purposes of monitoring change, this indicatorshould be measured at least every five years. Theindicator is not recommended for use at the locallevel as it may lead to unrealistic programexpectations. When interpreting trends inmortality, it is important not to combine indirectestimates with direct estimates in order to formtrends lines because of the possibility of spurioustrends, which may be an artifact of themethodologies used to estimate the rates anddifferences in the reporting periods.Sometimes, it is difficult to determine whether afetus has been born live or dead. Thisdetermination is often based on a judgment by anattending physician, nurse, or midwife and has an396influence of infant mortality measures. Ambiguityabout whether a live birth has occurred hasconsequences for whether the death is classifiedas a fetal death or a death to a live born.The indirect technique for calculating infant andchild mortality requires fewer and simplerquestions resulting in substantial cost savings.Collecting birth histories may be difficult and timeconsuming. However, information from birthhistories gives the most robust estimates of infantmortality, short of actual birth and deathregistration. Vital registration data may sufferfrom a number of problems, including omissionof some births and deaths, especially infants thatdied shortly after birth, and the misreporting ofthe date of birth and age at death. Omission ofinfant deaths is usually most severe for deaths thatoccur early in infancy.In developing countries, household surveys areessential to the calculation of this indicator, butthere are some limits to their quality. Survey dataare subject to recall error, and surveys estimatingunder-five deaths require large samples, whichmay be costly. Therefore, when using householdsurveys it is important to consider sampling errors.In addition, indirect estimates rely on estimatedactuarial (“life”) tables that may be inappropriatefor the population concerned.In settings with high AIDS prevalence, estimatesof under-five mortality using the direct or indirectapproach may be compromised. This potentialbias is due to the fact that children whose mothershave died tend to have higher child mortality, andin surveys, data on child deaths are collectedretrospectively from living mothers.Under five mortality rates based on householdsurveys can be calculated by different dimensionsof equity: sex of the child, geography (urban-ruralresidence, major regions, or provinces), or socioeconomicposition (mother’s education, wealthquintile). Often, disaggregated under-fivemortality rates are presented for ten-year periods,because of the rapid increase in sampling errors ifChapter 9

multiple categories are used. Censuses can alsoprovide such detail but are typically conducted everten years. Vital registration data do not often havethe socioeconomic background variables, but canprovide great geographic detail disaggregated byage and sex.The U5MR has been recommended as anindicator for determining the extent of vitamin Adeficiency (VAD) in a population. This is becausepopulations with high U5MR invariably tend tohave significant VAD. An U5MR > 50 per 1,000live births is considered to indicate a highlikelihood of VAD problems at the country level(IVACG, 2002). Under-five mortality rates arealso frequently used to compare levels ofsocioeconomic development across countries.Comparisons between countries should be madewith caution as the data used to arrive at estimatesof under-five mortality come from a wide varietyof sources of disparate quality and may refer todifferent reporting periods.Mortality397

-SCAUSEUSE-S-SPECIFICMORORTALITALITY RATEDefinitionThe mortality rate from a specified cause duringa given period.Numerator: Number of deaths attributed to aspecific cause in a given age group (e.g., 0-11months, 12-59 months) during a specifiedtime interval x 1,000 (or 10,000, 100,000).Denominator: Total population of that agegroup at the midpoint of the interval.Key illness groups for child survival include:• ARI/Pneumonia• Diarrhea• Febrile illness (Malaria)• Malnutrition• Neonatal tetanus• MeaslesMeasurement ToolsCommunity/demographic surveillance; vitalregistration; verbal autopsyWhat It MeasuresThis indicator measures deaths from specificcauses among children under the age of five years.Measurement of cause-specific mortality amongchildren is important for several reasons, includingthe following: (1) to establish the relative publichealth importance of the different causes of death;(2) to evaluate trends over time in the major killerdiseases; (3) to evaluate health interventions aimedat reducing mortality from specific causes of death,when these interventions are being introduced ina limited geographic area on a trial basis; (4) toinvestigate the circumstances surrounding thedeaths of children in order to identify ways toMortalityreduce unnecessary death; and (5) to facilitateresearch into factors associated with mortalityfrom specific causes of death.How to Measure ItIn calculating cause-specific mortality rates, thenumerator is the number of deaths attributed to aspecific cause during a specified time interval in apopulation, and the denominator is the size of thepopulation at the midpoint of the time interval.Cause-specific death rates can be calculatedseparately for children aged 0-11 months andthose aged 12-59 months.This indicator can be measured at the communityor health facility, district, and national levels. Inareas where medical certification of cause of deathis rare, verbal autopsy is used to identify the causesof death among infants and children.Demographic surveillance, where all deaths arereported on a regular basis throughout the year(often once every two weeks), can also be used foridentifying deaths.In verbal autopsy, two different methods can beused to obtain a verbal account of the cause ofdeath: an open-ended history of the final illnessand close-ended questions. In the open-endedhistory, the caregiver or next-of-kin is asked totell about the events leading up to the child’s deathin their own words, and then probed to follow-upon particular aspects. The descriptive account isthen reviewed by medical experts who code theinterview in terms of cause of death. Close-endedquestions ask whether specific symptoms and signswere present during the final illness.Strengths and LimitationsIt is difficult to assign causes of death.Misclassification of the cause of death not onlyaffects the estimated level of cause-specific399

mortality over time, but also differences in causespecificmortality rates between population groups.In mortality surveys, the accuracy of the indicatordepends on the ability of respondents to describethe final illness and on the way in which diseasesare understood and described in the community.At the community level, data collection is oftenretrospective and relies on verbal autopsy, ratherthan on a clinically-determined cause of death.These factors contribute to recall bias and makethe validity and reliability of the data questionable.Failure to enumerate all deaths can lead tomisleading results. In addition, very large samplesizes are required. In most sites, it will beimpossible to generate a cause-specific mortalityrate.Vital registration systems in developing countriesoften do not have sufficient coverage to provideaccurate data of cause-specific mortality.Demographic surveillance tends to cover limitedgeographic areas and underlying cause-specificmortality in these areas cannot necessarily begeneralized to wider populations.One limitation of this indicator is that the deathof a child is commonly the result of more thanone cause. Some verbal autopsy questionnaires,such as that developed by WHO, Johns HopkinsSchool of Hygiene and Public Health, and theLondon School of Hygiene and TropicalMedicine, allow for multiple causes of death, whileothers allow only one cause of death. Wheninterpreting this indicator, it is important to knowwhether multiple causes of death are allowed forin the coding, since the expected proportions ofdeath for each cause will generally be higher whenmultiple causes of death are allowed.400Chapter 9

RefererencesencesLawn J, McCarthy B, and Ross SR. (2001). The Healthy Newborn: A Reference Manual for Program Managers.CARE, Centers for Disease Control. CCHI and World Health Organization.Lumbignon P, Panamonta M, Laopaiboon M, Pothinam S, and Patithat N. (1990). Why Are Thai OfficialPerinatal and Infant Mortality Rates so Low? International Journal of Epidemiology, 19:997-1000.McCaw-Binns AM, Fox K, Foster-Williams KE, Ashley DE, and Irons B. (1996). Registration of Births,Stillbirths and Infant Deaths in Jamaica. International Journal of Epidemiology, 25:807-813.Rutstein SO. (1999). Guidelines for the MEASURE DHS+ Main Survey Report. Calverton, MD: MacroInternational, Inc. Demographic and Health Surveys.United Nations. (1983). Manual X: Indirect Techniques for Demographic Estimation. New York: UnitedNations.United Nations. (1994). Report of the International Conference on Population and Development, Programme ofAction of the International Conference on Population and Development, Cairo, Egypt, September 5-13, 1994.New York: United Nations. A/CONF. 171/13.United Nations. (2001). Principles and Recommendations for a Vital Statistics System Revision 2. New York:United Nations. Sales No. E.01.XV11.10.WHO. (1992). International Statistical Classification of Diseases and Related Health Problems, 10th Revision.Geneva, Switzerland: World Health Organization.WHO. (1996a). Mother-Baby Package: Implementing Safe Motherhood in Countries. Maternal Health andSafe Motherhood Program, Division of Family Health. Geneva, Switzerland: World Health Organization.WHO/FHE/MSM/94.11.WHO. (1999a). Indicators to Monitor Maternal Health Goals: Report of a Technical Working Group, Geneva,8-12 November 1993. Division of Family Health. Geneva, Switzerland: World Health Organization.WHO. (1999b). Reduction of Maternal Mortality: A Joint WHO, UNFPA, UNICEF, World Bank Statement.Geneva, Switzerland: World Health Organization.WHO and UNICEF. (1996). Revised 1990 Estimates of Maternal Mortality: A New Approach by WHO andUNICEF. Geneva, Switzerland: World Health Organization.WHO, UNICEF, and Wellstart International. (1999). The Baby Friendly Hospital Initiative: Monitoringand Reassessment: Tools to Sustain Progress. Geneva, Switzerland: World Health Organization.World Bank. (1993). 1993 World Development Report: Investing in Health. New York, NY: Oxford UniversityPress.Chapter 9. Mortality401