Anthropometric Indicators Measurement Guide

FOOD AND

NUTRITION

TECHNICAL

ASSISTANCE

AnthropometricIndicatorsMeasurementGuide

Bruce Cogill

SE

RIE

S

T I T L E 2

I N D I C ATO R G U I D E S

Bruce Cogill

AnthropometricIndicatorsMeasurementGuide

FOOD AND

NUTRITION

TECHNICAL

ASSISTANCE

Artwork proof 3 8/8/01 2:59 pm Page 1

This publication was made possiblethrough support provided by the Office ofHealth and Nutrition, Bureau for GlobalPrograms, US Agency for InternationalDevelopment, under the terms of AwardNo. HRN-A-00-98-00046-00, the Foodand Nutrition Technical AssistanceProject (FANTA). Additional support wasprovided by the Office of Food for Peace,Bureau for Humanitarian Response.Earlier drafts of the guide were developedwith funding from the Food and NutritionMonitoring Project (IMPACT) (ContractNo. DAN-5110-Q-00-0014-00, DeliveryOrder 16), managed by the InternationalScience and Technology Institute, Inc.and the Food Security Unit of theLINKAGES Project (CooperativeAgreement: HRN-A-00-97-00007-00),managed by the Academy for EducationalDevelopment. The opinions expressedherein are those of the author and do notnecessarily reflect the views of theUS Agency for International Development.It may be reproduced, if credit is given tothe FANTA Project.

Recommended citation: Cogill, Bruce.Anthropometric Indicators MeasurementGuide. Food and Nutrition TechnicalAssistance Project, Academy forEducational Development, WashingtonDC, 2001.

Published June 2001

Copies of the Guide can be obtained from:

Food and Nutrition Technical Assistance Project, Academy for Educational Development, 1825 Connecticut Ave., NW, Washington D.C., 20009-5721Tel: 202-884 8700. Fax: 202-884 8432. E-mail: [email protected] Website: www.fantaproject.org


8

10101011

12

1414

17172022

Page number Part 1.Introduction

Part 2.Anthropometric and annual monitoring indicators2.1 Anthropometric Indicators

2.1.1 Building blocks of anthropometry: indices2.1.2 What the indices reflect about the nutritional status of infants and children

2.2 Annual monitoring indicators

Part 3.Collecting anthropometric data through surveys3.1 Steps for conducting a survey

Part 4.Weighing and measuring equipment4.1 Scales4.2 Length/height boards4.3 Mid-upper arm circumference measure

Contents

1.

2.

3.

4.


Part 5.Taking measurements5.1 Interviewer field materials5.2 Procedures before measuring5.3 How to measure age, height, length, weight and MUAC

5.3.1 Age5.3.2 Height5.3.3 Length5.3.4 Weight5.3.5 MUAC

5.4 Assessing the accuracy of measurements5.5 Entering the data5.6 Training field staff

5.6.1 Planning the training5.6.2 Field exercises and standardization5.6.3 Survey training manual

Part 6.Comparison of anthropometric data to reference standards6.1 NCHS/WHO reference standards6.2 Comparisons to the reference standard6.3 Standard deviation units or z-scores6.4 Percentage of the median and percentiles6.5 Cut-offs

5.

6.

232324242426283034363637373838

393939404041

Contents - continued


7.

8.

9.

Part 7.Data analysis7.1 Source of Epi Info software7.2 Recommendations for analysis and presentation of height data7.3 Examples of data analysis7.4 Additional data analysis information

Part 8.Annual monitoring indicators8.1 Introduction8.2 Routine data collection8.3 Data on growth monitoring and promotion (GMP)

Part 9.ReferencesGlossaryAcronymsAppendix 1. Calculating Z-scoresAppendix 2. Uses of anthropometric dataAppendix 3. Selecting a sampleAppendix 4. Measuring adultsAppendix 5. Adolescent anthropometric indicatorsAppendix 6. Standardization of anthropometric measurementsAppendix 7. Guidelines for supervising surveysAppendix 8. Title II generic indicators

4344444548

49494950

5155636467697479829395


Acknowledgements

This guide draws extensively from materials from the AnthropometryResource Center, funded by the FAO/SADC project GCP/RAF/284/NET,Development of a Regional Food Security and Nutrition Information System,particularly the UN publication, How to Weigh and Measure Children:Assessing the Nutritional Status of Young Children in Household Surveys; andthe WHO publication, Physical Status: The Use and Interpretation ofAnthropometry. The training tool was developed by Bill Bender and SandyRemancus.

We thank the reviewers for their thoughtful comments during thedevelopment of this guide. The Cooperating Sponsors were essential to thedevelopment of the guide and this guide is dedicated to them. EunyongChung of the Global Bureau’s Office of Health and Nutrition has providedinsight and support for the guides and her efforts are appreciated. USAIDFood for Peace officers encouraged and supported the development of theguides. A number of people assisted in the development of the guide and thewriting of sections. Phil Harvey and Matthew Saaks rewrote and revisedsections of the guide and their input is greatly appreciated. SumathiSubramaniam and Laura Caulfield of Johns Hopkins University alsocontributed sections to the guide — Irwin Shorr, Penny Nestel, AnneSwindale, Patrick Diskin and Anne Ralte provided extensive comments andsupport for the guide.


This series of Title II Generic Indicator Guides has been developed by theFood and Nutrition Technical Assistance (FANTA) Project, and itspredecessor projects (LINKAGES and IMPACT), as part of USAID’s supportto develop monitoring and evaluation systems for use in Title II programs.These guides are intended to provide the technical basis for the indicatorsand the recommended method for collecting, analyzing and reporting on theindicators. A list of Title II Generic Indicators that were developed inconsultation with the cooperating sponsors in 1995/1996 is included inAppendix 8. The guides are available on the project websitewww.fantaproject.org.

Below is the list of available guides:Agricultural Productivity Indicators Measurement GuideFood Security Indicators and Framework for Use in the Monitoring and

Evaluation of Food Aid ProgramsInfant and Child Feeding Indicators Measurement GuideMeasuring Household Food Consumption: A Technical GuideSampling GuideWater and Sanitation Indicators Measurement Guide.

This series


1. IntroductionPA

RT

8

T his guide provides information on the Anthropometric ImpactIndicators and the Annual Monitoring Indicators for Maternal andChild Health/Child Survival (MCH/CS) and income-related Title II

activities, a subset of the P.L. 480 Title II Generic Performance Indicators forDevelopment Activities. The impact indicators are:

• decreased percent of stunted children (presented for ages 24-60months and by gender), where stunting is defined as percent of childrenfalling below -2 standard deviations for height-for-age;

• decreased percent of underweight children (in specified agegroupings such as 12-24 months 36-59 months and by gender) whereunderweight is defined as percent of children falling below -2 standarddeviations for weight-for-age.

These indicators are required for the reports of projects with specificnutrition components and are collected at baseline, mid-term and final-yearevaluations. Stunting, reflected by deficits in height-for-age would not beexpected to change in a short time period. It is recommended, therefore, notto report stunting figures annually. Underweight (or weight for age), reportedfor specific age groupings, would change more quickly as it is influenced byshort-term effects such as a recent outbreak of diarrheal diseases.

Some programs report stunting for children under 24 months of agerather than the recommended 24-60 months age grouping. Restricting theage grouping to children under 24 months has the disadvantage of notcapturing the lagged effects of the program and reducing the numbers ofpotential participants in a survey. The advantage of using children under 24months is that the data are more useful to determine the factors related tostunting for program design or redesign.The monitoring indicators are:

• increased percent of eligible children in growthmonitoring/promotion (usually presented for children <24 months or < 36 months of age, depending on the target group of the program);

1.


9

I N T RO D U C T I O N PA RT 1 .

• increased percent of children in growth promotion programgaining weight in past 3 months (by gender and age group, will dependupon the target group of the program).

The choice of indicators for annual monitoring and reporting shouldbe based upon a review of available sources of data and the informationneeds of the Cooperating Sponsor and USAID. Reporting the annualmonitoring indicators is recommended rather than required as in the casefor reporting on impact. The primary purpose of collecting and reportingthe monitoring indicators is to improve program management but theseindicators can also provide valuable insights into the interpretation of theanthropometric indicators of program impact. In addition, reporting theannual indicators may provide Cooperating Sponsors a further opportunityto demonstrate progress towards the achievement of results.

While the focus of this guide is on the consistent collection andreporting of nutritional anthropometry indicators and annualmonitoring indicators, suggestions are provided for additionalinformation related to monitoring and evaluation. This information willhelp Cooperating Sponsors to track and improve child nutrition activitiesand performance. The focus is on anthropometric assessment of infants andyoung children. The guide is a programming tool and is not meant tosubstitute for adequate technical and academic training needed to conductproblem analysis, design programs and for implementation. CooperatingSponsors are encouraged to seek technical expertise in nutritionalassessment and related topics needed to ensure appropriate useanthropometric indicators.

The assessment of children over 5 years of age, adolescents, adults andthe elderly is not the primary focus of the guide. Appendices 4 and 5, however,provide information on the nutritional assessments of adults and adolescents.

1.


2. Anthropometric Evaluation and AnnualMonitoring IndicatorsP

AR

T

10

Changes in body dimensions reflect theoverall health and welfare of individualsand populations. Anthropometry is used toassess and predict performance, healthand survival of individuals and reflect theeconomic and social well being of pop-ulations. Anthropometry is a widely used,inexpensive and non-invasive measure ofthe general nutritional status of anindividual or a population group. Recentstudies have demonstrated the applicationsof anthropometry to include the predictionof who will benefit from interventions,identifying social and economic inequityand evaluating responses to interventions.For more information on the application ofanthropometric data, refer to Appendix 2.

Anthropometry can be used for variouspurposes, depending on the anthrop-

ometric indicators selected. For example,weight-for-height (wasting) is useful forscreening children at risk and for meas-uring short-term changes in nutritionalstatus. However, weight-for-height is notappropriate for evaluating changes in apopulation over longer time periods. Aclear understanding of the different usesand interpretations of each anthropo-metric indicator will help to determine themost appropriate indicator(s) for programevaluation. For more detailed expla-nations of age and sex specific appropriateanthropometric uses, refer to Appendices4 and 5. Key terms are defined in theglossary.

2.

2.1. Anthropometric Indicators

The four building blocks or measures used to undertake anthropometric assessment are:

SEXAGELENGTH

(or height) WEIGHT

1 2 3 4

2.1.1.The Building Blocks of Anthropometry: Indices


11

I M PAC T I N D I C ATO R S F O R I M P ROV E D H O U S E H O L D N U T R I T I O N PA RT 2 .

Each of these variables provides one pieceof information about a person. When theyare used together they can provideimportant information about a person’snutritional status. The actual measure-ment of age, weight and height of childrenrequires specific equipment andtechniques which are described later.When two of these variables are usedtogether they are called an index. Threeindices are commonly used in assessingthe nutritional status of children:

• Weight-for-age;

• Length-for-age or Height-for-age;

• Weight-for-length or Weight-for-height.

There are many other anthropometricmeasures including mid-upper-arm-circumference (MUAC), sitting height tostanding height ratio (Cormic Index),and many skinfold measures. This guidewill concentrate on the measurementsand interpretation of weight and heightin children.

2.

2.1.2.What the Indices Reflect About the Nutritional Status of Infants and Children

The advantages and disadvantages of thethree indices and the information theycan provide is summarized below:

Weight-for-age: Low weight-for-ageindex identifies the condition of beingunderweight, for a specific age. Theadvantages of this index are that it mayreflect both past (chronic) and/or present(acute) undernutrition (although it isunable to distinguish between the two).

Height-for-age: This index is anindicator of past undernutrition orchronic malnutrition. It cannot measureshort term changes in malnutrition. Forchildren below 2 years of age, the term islength-for-age; above 2 years of age, theindex is referred to as height-for-age.Deficits in length-for-age or height-for-age are signs of stunting.

Weight-for-height: This index helps toidentify children suffering from currentor acute undernutrition or wasting and isuseful when exact ages are difficult todetermine. Weight-for-length (inchildren under 2 years of age) or weight-for-height (in children over 2 years of age)is appropriate for examining short-termeffects such as seasonal changes in foodsupply or short-term nutritional stressbrought about by illness.

The three indices are used to identifythree nutritional conditions: underweight,stunting and wasting.

Underweight: Underweight, based onweight-for-age, is a composite measure ofstunting and wasting and isrecommended as the indicator to assesschanges in the magnitude of malnutritionover time.

Stunting: Low length-for-age, stemmingfrom a slowing in the growth of the fetusand the child and resulting in a failure toachieve expected length as compared to ahealthy, well nourished child of the sameage, is a sign of stunting. Stunting is anindicator of past growth failure. It isassociated with a number of long-termfactors including chronic insufficientprotein and energy intake, frequentinfection, sustained inappro-priatefeeding practices and poverty. Inchildren over 2 years of age, the effects ofthese long-term factors may not bereversible. For evaluation purposes, it ispreferable to use children under 2 yearsof age because the prevalence of stuntingin children of this age is likely to be moreresponsive to the impact of interventionsthan in older children. Data onprevalence of stunting in a communitymay be used in problem analysis indesigning interventions. Information onstunting for individual children is usefulclinically as an aid to diagnosis.Stunting, based on height-for-age can beused for evaluation purposes but is notrecommended for monitoring as it doesnot change in the short term such as 6 -12 months.


2.

12

Wasting: Wasting is the result of a weightfalling significantly below the weightexpected of a child of the same length orheight. Wasting indicates current oracute malnutrition resulting from failureto gain weight or actual weight loss.Causes include inadequate food intake,incorrect feeding practices, disease, andinfection or, more frequently, a combi-nation of these factors. Wasting inindividual children and population groupscan change rapidly and shows markedseasonal patterns associated with changesin food availability or disease prevalenceto which it is very sensitive. Because of itsresponse to short-term influences, wastingis not used to evaluate Title II programsbut may be used for screening or targetingpurposes in emergency settings and issometimes used for annual reporting.Weight-for-height is not advised forevaluation of change in a population sinceit is highly susceptible to seasonality.

• Edema Edema is the presence ofexcessive amounts of fluid in theintracellular tissue. Edema can bediagnosed by applying moderate thumbpressure to the back of the foot or ankle.The impression of the thumb will remainfor some time when edema is present.Edema is diagnosed only if both feet show

the impression for some time. As aclinical sign of severe malnutrition, thepresence of edema should be recognizedwhen using short term indicators such aswasting. The presence of edema inindividuals should be recorded whenusing weight-for-height for surveillance orscreening purposes. When a child hasedema, it is automatically included withchildren counted as severely mal-nourished, independently of its wasting,stunting, or underweight status. This isdue to the strong association betweenedema and mortality. Edema is a rareevent and its diagnosis is used only forscreening and surveillance and not forevaluation purposes.

• Mid — Upper Arm Circumference(MUAC) MUAC is relatively easy tomeasure and a good predictor of immediaterisk of death. It is used for rapid screeningof acute malnutrition from the 6-59 monthage range (MUAC overestimates rates ofmalnutrition in the 6-12 month age group).MUAC can be used for screening inemergency situations but is not typicallyused for evaluation purposes. (MSF, 1995)MUAC is recommended for assessing acuteadult undernutrition and for estimatingprevalence of undernutrition at thepopulation level.

2.2. Annual Monitoring Indicators

Well chosen and reported monitoringindicators will enhance programmanagement and can provide valuableinsights into trends of anthropometricindicators used for determining impact.Part 8 of this Guide describes how annualmonitoring indicators that are based upondata from growth monitoring andpromotion programs (GMP) may becollected and reported in a standardformat. This is intended to make theindicators more useful for management ofprograms at all levels within countries,and also for reporting to USAID.

The two recommended annual monitoringindicators serve several purposes.

1. Percent of eligible children in GrowthMonitoring and Promotion programs

a. supports program management --providing information on coverage,targeting, and may provide a useful basisfor supervision of field staff;

b. provides information on context, orsome explanation, in the reporting ofanthropometric impact indicators; and

c. provides an indication of patterns of, ortrends in, service delivery and use andthus has potential to demonstratesuccesses of efforts to achieve specifiedproject results.


13

C O L L E C T I N G A N D A N A LY Z I N G T H E DATA PA RT 3 .

2. percent of children in GrowthMonitoring and Promotion (GMP) pro-grams gaining weight in past 3 months (by gender)

a. As a management tool, thisinformation is a trigger to increase growthpromotion and health educationcounseling. The information can be apositive communication between thehealth worker and caregiver concerningthe health of the child. This information ismost effective when provided with otherinformation such as food availability andpresence or history of infection.

b. As a surveillance tool, the indicatormay be useful as a lagged indicator of acommunity facing severe food or health-related stress. The usefulness of thisindicator for surveillance is reduced whensmall numbers are being monitored andaggregations over ages tend to even out .

A major advantage for an organization inreporting on the two monitoringindicators is that it provides national levelstaff with a framework to think about,interpret, and act upon data that arecurrently being reported to them. Abarrier to “institutionalizing” the report-ing of monitoring data is that often noaction is taken on information reportedand sometimes no meaningful feedback isprovided to the staff who collect and reportthem. Reporting on monitoring indicatorsat a national level will provide someevidence that GMP data have beencollected and used as intended.

Growth monitoring and promotionprograms are key components of manyfood assisted health and nutritionactivities of Title II programs. Thedevelopment of the guidelines presentedhere was based upon four assumptions:

a) growth monitoring without growthpromotion will not benefit the health ofparticipating children;

b) it is not useful to report village-levelactivities of GMPs to regional and nationallevels unless some action is taken, or somedecision is based upon the informationreported;

c) the reporting system will not besustainable without some meaningfulaction or feedback; and

d) continuing effort will be required toimprove the quality of data reported fromGMP activities, but reporting GMP datacan be useful when it is interpreted withinan appropriate context.

2.


3. CollectingAnthropometric DataThrough SurveysP

AR

T

14

The type of anthropometric data collectedwill depend on the reason for the survey.When the survey results will be used forlong-term planning the informationneeded might be different than inform-ation for program management. Theevaluation of Title II programs is asituation in which long-term changes instunting or undernutrition need to bereflected. Monitoring of growth promotionprograms will require different types ofinformation.

The collection of anthropometric data maybe the main purpose of a survey or it may bepart of a larger more comprehensive surveysuch as the KPC (Knowledge, Practice,Coverage). Information on individuals andhouseholds should be collected to interpretanthropometric data. Deciding whatinformation will be collected, how it will becollected and from whom it will becollected is all part of planning the survey.The steps that should be taken to conduct asurvey are outlined below.

3.

3.1. Steps for Conducting a Survey

The following checklist outlines the stepsnecessary for conducting a survey. All ofthese steps should be clearly thought outbefore the survey begins.

• Define survey objectives. The firststep is to determine the specific purpose ofthe survey. Make a detailed list of what isexpected to be achieved and whatinformation is needed.

• Budget for the survey. Develop adetailed item-by-item budget for all thecosts and expenses of the survey,including personnel, supplies, materials,transportation, accommodation andmeals. Determine the costs associatedwith data entry, cleaning, analysis,reporting and testing of all steps to ensuresmooth implementation.

• Choose the survey design. Dependingon the goal of the survey, the surveyplanning team should review differentdesign possibilities such as a case-controlor reflexive design before choosing thefinal design. Having a clear idea of thesurvey goals will help to determine whichpeople or which groups of people toinclude in the survey and the best methodfor collecting the information.

• Plan for personnel, facilities, andequipment. Conducting a survey within alimited time-frame (usually less than sixmonths) requires early planning formaterials and staff. During this stage thesurvey planning team decides how manyfield staff and how many office personnelthey will need and how they will recruitthem. Any advance work needed to find


3.

15

C O L L E C T I N G A N D A N A LY Z I N G T H E DATA PA RT 3 .

and hire staff is planned at this point.Other needs such as office space andequipment are also considered andplanned. Specific equipment is needed todo anthropometric assessments as part ofa survey and is discussed in Part 4.

• Select the sample. Once the surveygoals and methods of collecting theinformation have been decided, the groupsand numbers of people to be interviewedare selected. A sample is a small part ofthe group being studied that has beenchosen to represent the whole group.There are special considerations whenchoosing a sample for anthropometricassessment. Sampling is discussed inAppendix 3 and the reader is referredto the FANTA Sampling Guide.(www.fantaproject.org/publications)

• Develop the questionnaire. The list ofessential information needed to meet thesurvey objectives forms the basis of thesurvey questionnaire. A standard, printedquestionnaire ensures that all therespondents are asked the same questionsand enables the survey responses to betabulated easily and quickly. Thequestionnaire may need to be translatedinto local languages. Translatedquestionnaires should be translated backto the original language by anothertranslator and compared to the originalquestionnaire. Enumerators need to betrained in the appropriate use of thetranslation. See Figure 5.4 for a samplequestionnaire.

• Pre-test the questionnaire. Before thequestionnaire is finalized it should betested for content and length; thequestions should gather the neededinformation and should be easilyunderstood by both interviewers andrespondents. In the pre-test a smallnumber of interviews are conducted andthe questionnaire is revised on the basis ofthese results and comments from theinterviewers.

• Train personnel. Training of field staffis a vital step in the survey process;accurate, meaningful information can be

collected only if interviewers thoroughlyunderstand all their field instructions andprocedures. When all the field materialshave been prepared and finalized, and thefield staff has been hired, all interviewersand supervisors should be brought to acentral location to be taught surveyprocedures, how to collect the data andhow to use the questionnaire. Whenanthropometric assessment will be part ofthe survey, correct methods for takingmeasurements should also be part of thetraining schedule. If the actual survey isdelayed for more than three weeksfollowing training, it will be necessary toretrain personnel.

• Standardize the anthropometrictechnique. The training of personnel onspecific measurement and recordingtechniques should include not onlytheoretical explanations and demon-strations, but also provide an opportunityfor participants to practice the measure-ment techniques, as well as reading andrecording the results. Once all personnelhave adequately practiced the measure-ment and recording techniques, and feelcomfortable with their performance,standardization exercises should becarried out to ensure that all interviewersacquire the skills necessary to collect highquality data. Details of these exercises arepresented in Appendix 6.

• Interview. The success of a surveydepends on the quality of the fieldprocedures, supervision and interviewing.Interviewers should follow sampling andinterviewing instructions precisely andaccurately. They need to keep in touchwith their field supervisor and bring anyproblems or difficulties to their attention.

• Supervise the data collection. Oncethe interviewing begins, field supervisorsshould be present to assist interviewerswith problems that may arise in findingthe correct households, conducting theinterviews or completing the work ontime. Field supervisors, in addition tosolving field problems, are responsible fordistributing materials, reviewing andchecking completed questionnaires and


3.

16

making progress reports to the centraloffice. Detailed guidelines for supervisorsare presented in Appendix 7.

• Edit and code the interviews.Completed interviews should be reviewedto make sure all the questions have beenasked and the answers have beenrecorded clearly. Someone from thesurvey planning team should check allnumerical codes on the questionnaire andassign codes to any responses written inrespondents’ own words. Some surveysdirectly enter data into the computer at thetime of the measurement. This improvesquality and speed but requires functioningequipment in often difficult conditions.

• Tabulate the data. Whether thesurvey results are to be compiled by handor by computer, the responses for eachquestionnaire will have to be assignednumerical codes. This process is usuallysimplified by including numerical codesfor each of the response categories on tothe printed questionnaire form. When theinterviews are completed, these codes arethen transferred by hand to tabulationsheets or the codes can be entered into acomputer. Both of these methods allowthe survey results to be read andinterpreted by means of statistical tablesand percentages. As field computers andsatellite communication become morecommon, it will be possible to enter datadirectly by the interviewer withimmediate feedback for possible errors inmeasurement and recording.

• Analyze and report the survey results.On the basis of the tables prepared, thesurvey data are studied and interpretedand conclusions are drawn about thenutritional and socio-economic conditionsof the households in the project area. Thereport on the survey contains the mostimportant findings and conclusions,statistical tables and a description of theprocedures used in conducting the survey.The survey results should be presented ina clear and straightforward manner.


17

4.

4.Weighing and Measuring Equipment

PA

RT

Equipment is required to do anthropometricassessment. The most common types ofequipment used are scales and measuringboards. Sources for obtaining equipmentand tips on what to look for in equipmentare listed below. Whatever equipment ischosen, staff need training to ensure itsproper use and care. For Title II

Cooperating Sponsors, electronic weighingscales and locally adapted height measur-ing boards are recommended. Check withthe local UNICEF office for theirspecifications and availability. UNICEFequipment specifications can be found at:www.supply.unicef.dk/catalogue/index.htmin the 03 Nutrition Section.

4.1. Scales

Scales used in the field should be portable,durable and capable of reading up to 25 kgfor children and have 100 gramincrements. There are several differentattachments that can be used to helpweigh children with spring scales. Thesize of the child will determine whichattachment should be used. For weighinginfants, a sling or basket is usuallyattached to the spring scale. For children,weighing trousers are used to suspendthem. These are small pants with strapsthat the child steps into. The trousers arethen hung from the scale by the straps.There are other alternatives than thetrousers, but they can be difficult to use forinfants and small children. For infants, acloth folded to hang from the scale withthe infant is preferred. For children whoare old enough to grasp firmly ontosomething, a handle is sometimesattached to the scale and the child hangsfrom it by their hands until their weight isread. Whatever is used to suspend the

child, the scale should be zeroed to ensurethat the weight of the trousers, sling orbasket is not added to the child’s weight.

Oxfam Anthropometric Kit 1 (UNICEFItem No. 0000824) (Survey, screening,monitoring). The Anthropometric Kitcontains equipment for measuring theweight and height of children to assesstheir nutritional status, along with othermaterials for nutritional surveys. The kitweighs 26 kgs and contains measuringand survey materials for two surveyteams, or measuring equipment for 2feeding centers, and contains:

Oxfam Anthropometric Kit 1

Code Qty Description

NK3 2 Backpacks - nylon

NK6 2 Board - height/length & head block

NK7 4 Book - exercise

NK20 2 Calculator - solar and battery


Code Qty Description

NK26 4 Clipboard - A4, folding, spring clip

NK29 1 Roll cord 10M x 6mm, polypropylene,

endless fibres

NK30 2 Counter - manual/hand held, metal

NK31 4 Eraser

NK32 2 Forms - evaluation

NK33A 1 Notes on the revised Oxfam Feeding

Kits - English

NK33B 1 Notes on the revised Oxfam Feeding

Kits - French

NK34 40 Forms - survey

NK39 2 Pads, paper, four colors of paper,100

pages per pad, 90 x 90mm

NK40 1 Pad, paper, graph,A4 in mm

NK41 10 Pens - BIC ballpoint, black, medium

NK42 4 Pen - large, indelible, black marker

NK44 12 Pencil- HB

NK45 4 Pencil sharpener - metal, single hole

NK51 4 Ruler, 30cm, transparent, flat, plastic,

shatterproof

NK53 2 Scales, 25kg hanging scale, with bar, 3

pants and sling, 100gm graduation

NK55 1 Scissors - 17cm, blended

NK63 4 Tables - random number A4, plasticized

NK64 4 Tables - %Weight-for-height

NCHS/CDC/WHO Sex combined,

elasticized

NK65 4 Tables,Weight-for-height Z-score

NCHS/CDC/WHO Sex combined,

plasticized

NK66 2 Tape measure - fiberglass

NK67 50 Tapes, MUAC: red (<11cm),

orange (11-12.5cm),white 12.5-13.5cm),

green (>13.5cm)

NK68 2 Rolls, tape - metric adhesive

NK77 10 Wallets - A4 transparent plastic, open on

2-sides

Publications - Books

NK80 2 Food Scarcity and Famine - Oxfam

Practical Guide No. 7

NK83 1 MSF Nutrition Guidelines - English

NK84 1 MSF Nutrition Guidelines - French

NK85A 1 Refugee Health Care - Oxfam Practical

Guide No. 9 - English

NK85B 1 Refugee Health Care - Oxfam Practical

Guide No. 9 - French

NK86 1 Selective Feeding Program - Oxfam

Practical Guide 1 - English

NK87 1 Selective Feeding Program - Oxfam

Practical Guide 1 - French

4.

18

Unless weighing and measuringequipment is available locally, theAnthropometric Kit should be boughttogether with the Therapeutic Kit (Kit 4)or supplementary Kit (Kit 2 or Kit 3) forthe initial establishment of feedingprograms.

Electronic Scales: UNICEF and othershave found electronic scales to be durableand flexible especially given the option ofweighing the mother with the child. Themother can be weighed with the child.The mother is then weighed without thechild. The difference between the twomeasures is the child’s weight. Thistechnique is useful in situations where thechild struggles and use of a sling orweighing pants causes stress for the child.An additional advantage is that the weightof the mother is also available.

Each member of the field staff should havetheir own scale if possible, otherwise itmight take longer to do measurements andcomplete the survey. Several scales that areavailable for purchase are listed below:

UNICEF Electronic Scale (Item No.0141015 Scale mother/child, electronic)The scale is manufactured by SECA and isa floor scale for weighing children as wellas adults (capacity 150 kg). Weighingcapacity from 1 kg to 150 kg in 100 gdivisions, accuracy +/- 100 g. Weight ofadult on scale can be stored (tared) inmemory, allowing weight of baby or smallchild held by adult to show on scaleindicator. Solar cell on-switch (lightsensitivity 15 lux). Powered by long-lifelithium battery, good for one millionweighing cycles. Portable, weight 4 kg.Instructions in English, French and Spanish.

The major advantage of this scale is themicro-computer chip so that it can adjustto zero and weigh people quickly andaccurately. The child may be weigheddirectly. If a child is frightened, themother can first be weighed alone andthen weighed while holding the child inher arms, and the scale will automaticallycompute the child’s weight by subtraction.Recent experience in surveys suggests thatthe scale is appropriate for Cooperating


19

4.

Sponsor use although there have beensome difficulties with heat adverselyaffecting the scale. The price of this scaleis US$90. For more information contact:UNICEF Supply Division, CustomerServices and Support Center, UNICEFPlads, Freeport, DK-2000 Copenhagen,Denmark; Email:[email protected];Website: www.supply.unicef.dk. or contactUNICEF field office: www.unicef.org/uwwide/fo.htm.

UNICEF Hanging Scale (Item No.0145555 Scale, infant, spring, 25kgx100gwith No. 0189000 Weighing trousers/packof 5) This is a Salter type spring scale witha capacity of 25 kg and 100 gramincrements. Using this scale requires thatthe child be dressed in a set of plastic ornylon trousers before being weighed. Theinterviewers will need several pairs ofthese special trousers if they are going touse hanging scales. A hook for hangingthe scale from a door or a ceiling beammay also be necessary. The scale shouldbe checked periodically with standard 5 or10 kg weights. Beam-and-spring or dialtype scale, with two suspension hooks.With adjustment screw on top. ProvidedWITHOUT weighing trousers. Weighingtrousers must be ordered separately, in theproportion of 1 scale per 1 pack of 5trousers, item no. S0189000, in pack of 5.The scale weighs about 1 kg. For weightmonitoring. Practical to use, easy totransport. Suspend the scale from a solidsupport. The price of this scale is aboutUS$30. For more information contact:UNICEF Supply Division, CustomerServices and Support Center, UNICEFPlads, Freeport, DK-2000 Copenhagen,Denmark; Email:[email protected]; Website: www.supply.unicef.dk. or contactUNICEF field office: www.unicef.org/uwwide/fo.htm.

TALC Weighing Scale The TALC scalecan be used like any other hanging scale,with the advantage that a growth chart canbe put in it, and the child’s weight ismarked directly from the pointer on thespring. The TALC scale can be madelocally from a TALC starter kit. This

includes three springs, instructions andthree specimen growth charts. The scalecan also be made from local materials withthe purchase of the TALC scale spring andinstructions. A TALC sample pack includesspring, hook, nylon cord, wood pieces,screws and instructions. These can beordered from: Teaching Aids at Low Cost,P.O. Box 49, St. Albans, Herts. AL1 4AX,England. Payments from overseas must bemade by: (1) International Money Order,National Giro or U.K. Postal Order; (2)Sterling check drawn on a U.K. bank; (3)Eurocheque made out in Sterling; (4) USdollar check drawn on US bank using thecorrect rate of exchange; or (5) UNESCOCoupons.

Suspended Infant Weighing Pack -(Model No. PE-HS-25) This scale wasdeveloped in conjunction with the USCenters for Disease Control. It is a dialscale made of durable plastic with an easyto read face. It is capable of weighing upto 25 kg in 100 gram increments. Thepack includes a sling, weighing trousers,a detachable handle for weighing largerchildren and a vinyl shoulder bag. Theprice is US$150. Additional slings,trousers and handles are available forUS$12. For information on this packcontact: Perspective Enterprises, 7829Sprinkle Road, Kalamazoo, MI 49001,USA; Telephone: 800-323-7452; Fax: 616-327-0837.

Chasmors Ltd. Model MP25 This is alightweight scale with a stainless steelcase and an unbreakable plastic cover. Itis easy to read and can weigh up to 25kg in100 gram gradations. The scale comeswith two weighing trousers and one sling(for newborns). The price is US$75.Chasmors also carries a variety of armcircumference measuring tapes rangingfrom $US10 to US$30. For moreinformation on their products contact:CMS Weighing Equipment, Ltd., 18Camden High School, London NW1 OJH,U.K.; Telephone: 01-387-2060 (international)+44 171 387 2060.

W E I G H I N G A N D M E A S U R I N G E Q U I P M E N T PA RT 4 .


4.

20

Salter Model 235-6S This is a lightweightscale in a durable non-rust metal case withan unbreakable plastic face. Its capacity is25 kgs with 100g increments. The price isUS$77. For more information contact: Salter Industrial Measurement, Ltd.,George Street, West Bromwich, West Midlands, B70 6AD, U.K. Telephone: +44 121 553 1855. In the US - Salter Weighing Products 3620Central Avenue, N.E., MN 55418. Telephone: 1-800-637-0529. Email: [email protected]. www.salterbrecknell.com.

Medecins sans Frontieres NutritionSurvey Kit (MSF code: KMEDKNUT1--)This kit includes standard equipment forsurveys. It comes in one box and weighs21 kgs. The price is approximatelyUS$600. The kit can be bought at eitherOXFAM or TRANSFER (formerly associa-ted with MSF-Belgium). TRANSFER canbe contacted by email at [email protected] (state thatthe message is for TRANSFER).

4.2. Length/Height Boards

Length/height boards should be designedto measure children under 2 years of agelying down (recumbent), and olderchildren standing up. The board shouldmeasure up to 120 cm (1.2 meters) forchildren and be readable to 0.1 of acentimeter. A measuring board should belightweight, durable and have few movingparts. The metal part on the boardsabsorbs heat easily so care must be takenin field conditions. Another concern withlength/height boards is that they resemblecoffins and this can be disconcerting to thecaregiver. Check with the surveypersonnel and adjust the design. Provideadequate training both in using theequipment and in providing appropriateinformation for the caregiver. Ideally,each field staff should have their ownboard. This makes the survey processmore efficient than when boards have tobe shared. Several types of length andheight boards are available and are listedbelow. The Dutch infant-child-adultmeasuring board is recommended butlocal adaptations are possible to reducethe cost.

UNICEF Model (Item No. 0114500 Infantlength/height measuring board) Aninfant/child height measuring boardmeasuring both recumbent length andstanding height. This board is made ofwood, smooth-finish, all parts glued andscrewed; height is 130 cm (collapses to 75cm); and width 30 cm. Supplied with a

shoulder strap. Illustrated instructions forassembly and use are included, alsoguidelines and plans for localconstruction. Estimated weight: 10 kg.The price of the board is about US$350.For more information contact: UNICEF Supply Division, CustomerServices and Support Center,

UNICEF Plads, Freeport, DK-2000 Copenhagen, Denmark;Email:[email protected]; Website: www.supply.unicef.dk. or contactUNICEF field office: www.unicef.org/uwwide/fo.htm.

Infant/Child Height/Length MeasuringBoard This board has 130 cm capacity(collapses to 75cm) and has 0.1 cmincrements. The board weighs 6 kg, isportable, water-resistant and has anadjustable, removable nylon shoulderstrap. It is easy to assemble and dismantle,with the sliding head-footpiece stored in thebase of the board for transport or storage.This board has a lifetime warranty andcosts $285. For more information contact: Shorr Productions, 17802 Shotley Bridge Place, Olney, Maryland 20832, USA;Telephone: 301-774-9006 (toll free 877-900-9007); Fax: 301-774-0436; Email: [email protected].

Infant Recumbent Length Board (Model No. PE-RILB-122-PC) This board is


21

4.

lightweight, durable and capable ofmeasuring recumbent length up to 100cm. The price of this board is US$215.For more information contact: Perspective Enterprises, 7829 Sprinkle Road, Kalamazoo, MI 49001, USA; Telephone: 800-323-7452; Fax: 616-327-0837.

Recumbent Length Measuring BoardThis board has 106 cm capacity in 0.1cm increments and costs $185. Formore information contact: Shorr Productions, 17802 Shotley Bridge Place, Olney, Maryland 20832, USA;Telephone: 301-774-9006

(toll free 877-900-9007); Fax: 301-774-0436; Email: [email protected].

Infant Measuring Board (Model No. PE-RILB-LTWT) This measuring board isdesigned to be especially lightweight andextremely portable. It can measure up to100 cm, is collapsible and comes with avinyl plastic tote bag. The price of thisboard is US$250. For more informationcontact: Perspective Enterprises, 7829 Sprinkle Road, Kalamazoo, MI 49001, USA; Telephone: 800-323-7452; Fax: 616-327-0837.

Portable Adult/Infant Measuring Board(Model No. PE-AIM-101) This is anadjustable measuring board which hasbeen extensively used by WHO and CDC,with a vertical aluminum post. It canmeasure the height of adults and then bytaking off its vertical extension it can beadapted to measuring infants. Whencollapsed it is approximately the length oftwo briefcases laid end to end. It also hasan optional vinyl carrying case. The priceof this board is US$385. For moreinformation contact: Perspective Enterprises, 7829 Sprinkle Road, Kalamazoo, MI 49001, USA.; Telephone: 800-323-7452; Fax: 616-327-0837.

Infant/Child/Adult Height/LengthMeasuring Board This measuring boardhas a 200 cm capacity (collapses to 85 cm)and has 0.1 cm increments. The boardweighs 6 kg, is portable, water-resistantand has an adjustable, removable nylonshoulder strap. It is easy to assemble anddismantle and costs $285. For moreinformation contact: Shorr Productions, 17802 Shotley Bridge Place, Olney, Maryland 20832, USA; Telephone: 301-774-9006 (toll free 877-900-9007); Fax: 301-774-0436; Email: [email protected].

Adult Measuring Device (Microtoise)(UNICEF No. 0114400 Height measuringinstrument (0-2 m)) This lightweightportable tape is wall mounted and fitseasily into the package needed for fieldmeasurements. Made of plastic, theMicrotoise measures up to 2 meters and isavailable at a price of approximately$US20. For more information contact:UNICEF Supply Division, CustomerServices and Support Center,

UNICEF Plads, Freeport, DK-2000 Copenhagen, Denmark; Email:[email protected]; Website: www.supply.unicef.dk. or contactUNICEF field office: www.unicef.org/uwwide/fo.htm.

Adult Measuring Device (HarpendenPocket Stadiometer) An inexpensive heightmeasuring device useful for children over24 months and adults (range 0-2000mm).The price is approximately $US100. CMS Weighing Equipment, Ltd., 18 Camden High School, London NW1 OJH, UK; Telephone: 01-387-2060 +44 171 387 2060.

Local Construction Various plans existfor the local construction of foldableheight/length boards and they can bemade for around $US20. It is importantthat the materials are durable, lightweightand the wood should be well seasoned toguard against warping. Sealing the woodwith water repellant and ensuring themeasuring tape is protected from wear

W E I G H I N G A N D M E A S U R I N G E Q U I P M E N T PA RT 4 .


4.

22

will improve the durability of the board.The tape measure should be durable with0.1 cm increments and the numbers of thetape measure must be next to the markingson the board when the measure is glued tothe side of the board. The boards shouldbe long enough to measure children up to5 years and a “correction” factor is neededto convert recumbent length to standingheight for children over 24 months inorder to use the WHO/NCHS growth

reference standards. Designs can be foundin Annex 4 of the FAO field manual.Blueprints for the construction of portablemeasuring boards are available from theNutrition Division of Cornell University aswell as from the Center for HealthPromotion and Education:Centers for Disease Control and Prevention,1600 Clifton Road, N.E., Atlanta, GA 30333,USA.

4.3. Mid-Upper Arm Circumference Measure

MUAC Tape (UNICEF Item No. 145600Arm circumference insertion tape/pack of50) Arm circumference insertion tape: tomeasure mid-upper arm circumference ofchildren, up to 25 cm. Color-coded inred/yellow/green, non-tear, stretch-resistant plasticized paper. Supplied inpack of tapes together with written andpictorial instructions for use. Refer toUNICEF's Supply Division in Copenhagen

through any UNICEF field office. For moreinformation contact: UNICEF Supply Division, CustomerServices and Support Center, UNICEF Plads, Freeport, DK-2000 Copenhagen, Denmark;Email:[email protected]; Website: www.supply.unicef.dk. or contact UNICEF field office:www.unicef.org/uwwide/fo.htm.


23

5.

5.Taking Measurements

PA

RT

Accurate anthropometric measurement is askill requiring specific training. A numberof tools are available and additionalreferences and sources are provided inAppendix 6. Training requires step-by-stepprocedures to follow when takingmeasurements. Standardizing methodshelps ensure that the measurements will becorrect and makes comparisons possible.

Comparisons may be done between datacollected from different areas of a country,between different surveys or betweenmeasurements and the reference standards.None of these comparisons will be possiblewithout a standard method for takingmeasurements. This section will cover thenecessary field equipment and methods fortaking measurements.

5.1. Interviewer field materials

The checklist below includes the equipmentand materials interviewers should have withthem in the field. All of these items may notbe necessary for every survey.

• Equipment bag

• List of assigned households and their addresses (or location)

• Map of the area

• Log book

• Pre-numbered questionnaires for assigned households

• Spare questionnaires

• Waterproof envelopes for blank and completed questionnaires

• Weighing scale

• Scale hooks

• Weighing pants or hanging swing

• Storage bag for pants

• Piece of rope for scales

• Storage box for scales

• Height/length measuring board

• Sliding head/foot pieces

• Clipboard

• Stapler and box of staples

• Pencils and pencil sharpener

• Eraser

• Pens

• Spare paper


5.

24

5.2. Procedures Before Measuring

There are a few preparatory proceduresand decisions that should be addressedprior to obtaining measurements.Guidelines to make the field experienceeasier are:

• Initial preparation. Ensure that themother or caretaker understands what ishappening to the child. The measurementof weight and length can be traumatic.Participants need to be comfortable withthe process. The equipment should becool, clean and safely secured. Work outof direct sunlight since it can interferewith reading scales and other equipmentand it is more comfortable for themeasurer and child.

• Two trained people required.When possible, two trained people shouldmeasure a child’s height and length. Themeasurer holds the child and takes themeasurements. The assistant helps holdthe child and records the measurementson the questionnaire. If only one trainedperson is available to take themeasurements, then the child’s mothercan help. The measurer would alsorecord the measurements on thequestionnaire.

• Measuring board and scaleplacement. There will usually be severalchoices on where to place the measuring

board or scale, but the choice should bemade carefully. Be sure that you have asturdy, flat surface for measuring boards,a strong place to hang scales from andadequate light so the measurements canbe read with precision.

• When to weigh and measure.Weighing and measuring should not bethe first thing you do when you start aninterview. It is better to begin withquestions that need to be answered. Thishelps make the mother and child feelmore comfortable before the measure-ments begin.

• Weigh and measure one child at atime. You should complete the questionsand measurements for one child at a time.This avoids potential problems with mix-ups that might occur if you have severalchildren to measure.

• Control the child. When you aretaking weight and length/heightmeasurements the child needs to be ascalm as possible. A child who is excitedor scared can make it difficult to get anaccurate measurement.

• Recording measurements. Allmeasurements should be recorded inpencil. If a mistake is made when recordinga measurement, it can be corrected.

5.3. How to Measure Age, Height, Length,Weight and MUAC

The following suggestions are adaptedfrom How to Weigh and Measure Children:Assessing the Nutritional Status of YoungChildren in Household Surveys, UNDepartment of Technical Cooperation forDevelopment and Statistical Office, 1986.

5.3.1 AgeThe child’s accurate age is required forsampling, deciding on whether the child ismeasured standing or reclining for heightor length, and for converting height andweight into the standard indices. At thetime of measurement, an age estimate isneeded for decisions on sampling and for

the position on the measuring board. It isrecommended the enumerators use simplemethods to approximate the age and thatthe data analyst calculates the age using acomputer program which will require thedate of birth and date of measurement.

To complete the determination, theenumerator needs to examine documen-tary evidence of the birth date (such asbirth, baptismal certificate, clinic care orhoroscope). Cross checking is necessaryeven if the mother knows the birth date orage of the child as errors in recall arecommon. Where there is a generalregistration of births and where ages are


25

5.

generally known, the recording of age is astraightforward procedure, with agemeasured to the nearest month or year asthe case may be. For example, an infantwhose date of birth is 13 July, 1996 couldbe recorded as being 6 months if seenbetween 13 December, 1996 and 12January, 1997 (both dates inclusive).Similarly, a child born on 13 July, 1995could be recorded as 6 years old if seenbetween 13 July 2001 and 12 July, 2002(both dates inclusive). If dates cannot berecalled, use of a local calendar will assistmothers in recalling the date of birth.Construction of the local calendar shouldbe done prior to the survey and testedusing the enumerators.

TA K I N G M E A S U R E M E N T S PA RT 5 .


5.

26

5.3.2. Height for children 24 months andolder (Figure 5.1)

1. Measurer or assistant: Place themeasuring board on a hard flat surfaceagainst a wall, table, tree, staircase, etc.Make sure the board is not moving.

2. Measurer or assistant: Ask themother to remove the child’s shoes andunbraid any hair that would interferewith the height measurement. Ask her towalk the child to the board and to kneel infront of the child. If a Microtoise measureis used, stand the child vertically in themiddle of the platform.

3. Assistant: Place the questionnaire andpencil on the ground (Arrow 1). Kneelwith both knees on the right side of thechild (Arrow 2).

4. Measurer: Kneel on your right kneeon the child’s left side (Arrow 3). This willgive you maximum mobility.

5. Assistant: Place the child’s feet flat andtogether in the center of and against theback and base of the board/wall. Place yourright hand just above the child’s ankles onthe shins (Arrow 4), your left hand on thechild’s knees (Arrow 5) and push againstthe board/wall. Make sure the child’s legsare straight and the heels and calves areagainst the board/wall (Arrows 6 and 7).Tell the measurer when you havecompleted positioning the feet and legs.

6. Measurer: Tell the child to lookstraight ahead at the mother who shouldstand in front of the child. Make sure thechild’s line of sight is level with theground (Arrow 8). Place your open lefthand under the child’s chin. Graduallyclose your hand (Arrow 9). Do not coverthe child’s mouth or ears. Make sure theshoulders are level (Arrow 10), the handsare at the child’s side (Arrow 11), and thehead, shoulder blades and buttocks areagainst the board/wall (Arrows 12, 13,and 14). With your right hand, lower theheadpiece on top of the child’s head.Make sure you push through the child’shair (Arrow 15).

7. Measurer and assistant: Check thechild’s position (Arrows 1-15). Repeat anysteps as necessary.

8. Measurer: When the child’s positionis correct, read and call out themeasurement to the nearest 0.1 cm.Remove the headpiece from the child’shead and your left hand from the child’schin.

9. Assistant: Immediately record themeasurement and show it to themeasurer.

10. Measurer: Check the recordedmeasurement on the questionnaire foraccuracy and legibility. Instruct theassistant to erase and correct any errors.


27

5.

Figure 5.1. Child Height Measurement - Height for Children 24 Months and Older

Source: How to Weigh and Measure Children: Assessing the Nutritional Status of Young Children, UN 1986.

4

Right hand on shins; heels against back and base of board

Headpiece firmly on head

9

12

13

14

6

7

Hand on chin

10Shoulders level

11

15

Child's handsand arms at side

5Left hand onknees; knees

togetheragainst board

8

Lineof sight

3

Measurer on knees

2

1

Assistant on knees

Questionaire and pencil onclipboard on floor or ground



5.

28

5.3.3. Length for infants and children 0-23 months (Figure 5.2)

1. Measurer or assistant: Place themeasuring board on a hard flat surface,i.e., ground, floor, or steady table.

2. Assistant: Place the questionnaire andpencil on the ground, floor, or table(Arrow 1). Kneel with both knees behindthe base of the board if it is on the groundor floor (Arrow 2).

3. Measurer: Kneel on the right side ofthe child so that you can hold the footpiece with your right hand (Arrow 3).

4. Measurer and assistant: With themother’s help, lay the child on the boardby supporting the back of the child’s headwith one hand and the trunk of the bodywith the other hand. Gradually lower thechild onto the board.

5. Measurer or assistant: Ask themother to kneel close on the opposite sideof the board facing the measurer as thiswill help to keep the child calm.

6. Assistant: Cup your hands over thechild’s ears (Arrow 4). With your armscomfortably straight (Arrow 5), place thechild’s head against the base of the boardso that the child is looking straight up.The child’s line of sight should beperpendicular to the ground (Arrow 6).Your head should be straight over thechild’s head. Look directly into the child’seyes.

7. Measurer: Make sure the child islying flat and in the center of the board(Arrows 7). Place your left hand on thechild’s shins (above the ankles) or on theknees (Arrow 8). Press them firmlyagainst the board. With your right hand,place the foot piece firmly against thechild’s heels (Arrow 9).

8. Measurer and assistant: Check thechild’s position (Arrows 1-9). Repeat anysteps as necessary.

9. Measurer: When the child’s positionis correct, read and call out themeasurement to the nearest 0.1 cm.Remove the foot piece and release your lefthand from the child’s shins or knees.

10. Assistant: Immediately release thechild’s head, record the measurement, andshow it to the measurer.



29

5.

3

Measurer on knees

2

Assistant on knees

Child flat on boardChild's feet flatagainst footpiece

8

Questionaire and pencil onclipboard on floor or ground

Hand on knees orshins; legs straight

5Arms comfortably straight

79

Hands cupped over ears;head against base of board

1

4

6Line of sight perpendicular to base of board

90o

Figure 5.2. Child length measurement - length for infants andchildren 0-23 months




5.

30

5.3.4. Weight Using Salter-like HangingScale (Figure 5.3)

1. Measurer or assistant: Hang thescale from a secure place like the ceilingbeam. You may need a piece of rope tohang the scale at eye level. Ask the motherto undress the child as much as possible.

2. Measurer: Attach a pair of the emptyweighing pants to the hook of the scaleand adjust the scale to zero, then removefrom the scale.

3. Measurer: Have the mother hold thechild. Put your arms through the leg holesof the pants (Arrow 1). Grasp the child’sfeet and pull the legs through the leg holes(Arrow 2). Make certain the strap of thepants is in front of the child.

4. Measurer: Attach the strap of thepants to the hook of the scale. DO NOTCARRY THE CHILD BY THE STRAPONLY. Gently lower the child and allowthe child to hang freely (Arrow 3).

5. Assistant: Stand behind and to oneside of the measurer ready to record themeasurement. Have the questionnaireready (Arrow 4).

6. Measurer and assistant: Check thechild’s position. Make sure the child ishanging freely and not touching anything.Repeat any steps as necessary.

7. Measurer: Hold the scale and read theweight to the nearest 0.1 kg. (Arrow 5).Call out the measurement when the childis still and the scale needle is stationary.Even children who are very active, whichcauses the needle to wobble greatly, willbecome still long enough to take areading. WAIT FOR THE NEEDLE TOSTOP MOVING.

8. Assistant: Immediately record themeasurement and show it to themeasurer.

9. Measurer: As the assistant records themeasurement, gently lift the child by thebody. DO NOT LIFT THE CHILD BY THE

STRAP OF THE WEIGHING PANTS.Release the strap from the hook of thescale.



31

5.

Figure 5.3. Child Weight


1

5

4

2

3 Child hangs freely

Measurer reads scale at eye level

Assistant withquestionaire

Put hands throughlegholes

Grasp feet



5.

32

5.3.4. Child Weight Using UNICEFUNISCALE (Figure 5.4)

The UNICEF electronic scale requires themother and child to be weighedsimultaneously. Minimize the clothing onthe child. Ensure the scale is not over-heated in the sun and is on an evensurface enabling the reading to be clear.Ask the mother to stand on the scale.Record the weight and include the readingwith one decimal point (e.g. 65.5 kgs).Pass the child to a person nearby. Recordthe second reading with just the mother(e.g. 58.3 kgs). The difference (e.g. 7.2 kgs)is the weight of the child. Refer to theUNICEF document “How to Use theUNISCALE” (June, 2000) prepared by theNutrition Section Program Division/UNICEF New York.Contact:Esther Liestyowati, Nutrition Section,UNICEF New York, fax: 212.824.6465, e-mail: [email protected]


33

5.

Figure 5.4. Child Weight Measurement

Source: "How to use the UNISCALE" UNICEF, 2000



5.

34

5.3.5. Child Mid-Upper Arm Circum-ference (MUAC) Procedure (Figure 5.5)

1. Measurer: Keep your work at eye level.Sit down when possible. Very youngchildren can be held by their motherduring this procedure. Ask the mother toremove clothing that may cover the child’sleft arm.

2. Measurer: Calculate the midpoint ofthe child’s left upper arm by first locatingthe tip of the child’s shoulder (Arrows 1and 2) with your finger tips. Bend thechild’s elbow to make a right angle (Arrow3). Place the tape at zero, which isindicated by two arrows, on the tip of theshoulder (Arrow 4) and pull the tapestraight down past the tip of the elbow(Arrow 5). Read the number at the tip ofthe elbow to the nearest centimeter.Divide this number by two to estimate themidpoint. As an alternative, bend the tapeup to the middle length to estimate themidpoint. A piece of string can also beused for this purpose. Either you or anassistant can mark the midpoint with apen on the arm (Arrow 6).

3. Measurer: Straighten the child’s armand wrap the tape around the arm atmidpoint. Make sure the numbers areright side up. Make sure the tape is flataround the skin (Arrow7).

4. Measurer and assistant: Inspect thetension of the tape on the child’s arm.Make sure the tape has the proper tension(Arrow 7) and is not too tight or too loose(Arrows 8-9). Repeat any steps asnecessary.

5. Assistant: Have the questionnaireready.

6. Measurer: When the tape is in thecorrect position on the arm with thecorrect tension, read and call out themeasurement to the nearest 0.1cm.(Arrow 10).

7. Assistant: Immediately record themeasurement on the questionnaire andshow it to the measurer.

8. Measurer: While the assistant recordsthe measurement, loosen the tape on thechild’s arm.


10. Measurer: Remove the tape from thechild’s arm.


35

5.

6 8 9 10 11 12 13 14 15 16 18

9 10

19 20 21 22 23 24 257cm

0. cmArm circumference “insertion” tape

0. cm

2 4

6

53

7

8

9

1

10

Figure 5.5. Child Mid-Upper Arm Circumference Measurement

4. Place tape at tip of shoulder5. Pull tape past tip of bent elbow

6. Mark midpoint2. Tip of shoulder3. Tip of elbow

1. Locate tip of shoulder

7. Correct tape tension

8. Tape too tight

9. Tape too loose 10. Correct tape position for arm circumference

How to weigh and measure children: assessing the nutritional status of young children in UN (1986)



5.

36

5.4.Assessing the Accuracy of Measurements

Accuracy is achieved through goodtraining and supervision. There aretechniques for measuring the accuracy ofthe measurements. When taking morethan one height or weight measurementon the same person, the twomeasurements can be averaged. If theyare vastly different from each other, themeasurements should be disregarded andthe measuring should start again (Table5.1 provides specific parameters).

Table 5.1. Largest acceptable differences between repeated measurements

Anthropometric Largest acceptablemeasurement difference

Weight 0.5kg

Height 1.0cm

MUAC 0.5cm

These are:

• Checking the measurements recordedand submitted by field staff, to see whetherthey look reasonable.

• Accompanying field staff on interviewsto watch how measurements are taken.

• Conducting repeat visits to somehouseholds that have already been inter-viewed by the field staff. Measurementsshould be repeated to determine if theprevious measurements are supported bythe repeat measurements.

Appendix 6 has a section on AnthropometricStandardization tests. These tests can beused during training or at any point duringthe survey process to check how accuratelyfield staff take measurements. These testscan be especially useful during training todetermine who needs more training or whomight need a little extra supervision once inthe field.The field supervisor is usually responsible

for assessing the accuracy ofmeasurements. There are a few practicesa supervisor should employ to make surethat the data collected is of high quality.

5.5. Entering the data

A survey questionnaire usually contains awide range of information to be collected.A questionnaire should be adapted to theneeds for measuring anthropometry.Some information will carry over fromone module or section to another. Thefollowing is an example of one formatused for survey work for children underfive years of age.

Anthropometry - basic information:Enter the children’s names andidentification code numbers, enter the sexand their ages (see Figure 5.6). Be carefulnot to mix up children when moving fromone section of a questionnaire to another.

Child weight: Record the child’s weightin kilograms to one decimal. Read thesupporting notes carefully as they shouldbe known to all interviewers andsupervisors. In the example of Mary

(Figure 5.6), her weight was 10.2kilograms. Had her weight been 9.5 kgs,the entry would be |0|9:5|. Always note thezeros and the decimals.

Child length: Record the child’s lengthin centimeters to one decimal. In theexample of Mary (Figure 5.6), her lengthwas 67.3 centimeters. The entry is madeas |0|6|7:3|. Always note the zeros and thedecimals. Make sure the information isentered accurately and fully on each child.


37

5.

Figure 5.6. Child anthropometry questionnaire (partial)

Section 1: Health and Nutrition

1.1 Anthropometry

Household ID

Date of interview

NameChildID

Date of birth(dd/mm/yy)

Age(months)

Sex1 = Male2 = Female

Weight(kg 0.1)

To the nearest0.1kg and 0.1cm

For Example:Mary, a girl, born on 7 August 1996, is approximately 11 months old and weighed 10.2 kilograms and was 67.3 centimetres long

Mary

Length(cms 0.1)

dd mm yy

dd mm yy

0 1

0 1

dd mm yy

0 2

dd mm yy

0 0 7 0 8 9 6 1 1 2 2 31 0 60 72 : :

: :

: :

5.6.Training field staff

Training field staff to collect anthropo-metric data through surveys usuallyinvolves learning: to take anthropometricmeasurements; and other skills such ashousehold selection, interviewingtechniques and recording requirements.All of these skills are important forconducting surveys that yield valid results.This section will cover what should beexpected from field staff training.

5.6.1 Planning the trainingIt is recommended that you always selectmore candidates than you need. This willallow you to pick the candidates with thebest performance when training is overand will give you some extra trainees incase of dropouts.

The length of the training will varydepending on the resources available andthe complexity of the survey. As a

guideline, training is generally scheduledfor two to five days. Usually, the first dayof training is spent explaining the purposeof the survey and outlining the surveyprocedures; the second and third daysfocus more closely on survey proceduresand the questionnaire; and the last coupleof days should be used for field exercisesand tests. Field exercises will be coveredin more detail later in this section.

The checklist below lists the topics thatshould be covered during training:

• Purpose and background of the survey

• Organization of the survey team anddivision of responsibilities

• Explanation of sampling andhousehold selection procedures

• Question-by-question review of thequestionnaire

• Instruction in techniques of inter-viewing, recording answers and checking



5.

38

out questionnaires

• Explanation of specific nutritionindicators

• Instructions on how to take and recordanthropometric measurements andstandardization tests

• Administrative details (timetable, log-book, supplies, reports).

5.6.2. Field exercises and standardizationSurvey staff should have ampleopportunity to practice the skills taughtduring training. This is especially truewith training on taking anthropometricmeasurements. Trainees practical skillsneed to be developed. During practicesessions a supervisor can determine whoneeds more training. Practice sessionsmight begin by taking trainees to a school,maternal and child health clinic, hospitalor orphanage and letting them practicetaking children’s measurements. Thestandardization exercises described inAppendix 6 ensure the trainees haveacquired the necessary skills.

By the end of training, all traineesshould also have had a chance to practicewhat they have learned. Choose a villagethat is close to the training center. Thetrainees should go through an entiresurvey with a few households and thesupervisor should watch how each traineeperforms. This will provide trainees withhands-on experience, make them feelmore confident when they go into the fieldand will give the supervisor a chance tocorrect any mistakes.

5.6.3. Survey training manualA training manual should give an overviewof the purpose of the survey, an outline ofthe whole survey process and clearlydefine what is expected of the field staff. Itcan also include useful tips and answers tocommon questions that come up in thefield. All field staff should be provided withtheir own copy of the training manual.


6.

39

6.Comparison ofAnthropometric Data to

Reference Standards

PA

RT

Comparing the measurements of children toreference standards is an easy procedurebecause of readily available public-domaincomputer software. This section describes

some underlying principles for efficient useof the available software beginning withhow individual measurements arecompared to the reference standard.

6.1. NCHS/WHO Reference Standards

The reference standards most commonlyused to standardize measurements weredeveloped by the US National Center forHealth Statistics (NCHS) and arerecommended for international use by theWorld Health Organization. The referencepopulation chosen by NCHS was astatistically valid random population ofhealthy infants and children. Questionshave frequently been raised about thevalidity of the US-based NCHS referencestandards for populations from other ethnicbackgrounds. Available evidence suggests

that until the age of approximately 10 years,children from well-nourished and healthyfamilies throughout the world grow atapproximately the same rate and attain thesame height and weight as children fromindustrialized countries. The NCHS/WHOreference standards are available forchildren up to 18 years old but are mostaccurate when limited to use with childrenup to the age of 10 years. The internationalreference standards can be used forstandardizing anthropometric data fromaround the world.

6.2. Comparisons to the Reference Standard

References are used to standardize a child’smeasurement by comparing the child’smeasurement with the median or averagemeasure for children at the same age andsex. For example, if the length of a 3 monthold boy is 57 cm, it would be difficult to knowif that was reflective of a healthy 3 month oldboy without comparison to a referencestandard. The reference or median lengthfor a population of 3 month old boys is 61.1cm and the simple comparison of lengths

would conclude that the child was almost 4cm shorter than could be expected.

When describing the differences fromthe reference, a numeric value can bestandardized to enable children of differentages and sexes to be compared. Using theexample above, the boy is 4 cm shorter thanthe reference child but this does not take theage or the sex of the child into consideration.Comparing a 4 cm difference from thereference for a child 3 months old is not the


40

6.

same as a 4 cm difference from thereference for a 9 year old child, because oftheir relatively different body sizes.

Taking age and sex into consideration,differences in measurements can beexpressed a number of ways:

• standard deviation units, or Z-scores

• percentage of the median

• percentiles.

To standardize reporting, USAIDrecommends that Cooperating Sponsorscalculate percentages of children below cut-offs as well as other statistics using Z-scores.If Z-scores cannot be used, percentage of themedian should be used.

6.3. Standard Deviation Units or Z-Scores

Z-scores are more commonly used by theinternational nutrition community becausethey offer two major advantages. First,using Z-scores allows us to identify a fixedpoint in the distributions of differentindices and across different ages, i.e., for allindices for all ages, 2.28% of the referencepopulation lie below a cut-off of -2 Z-scores. The percent of the median does nothave this characteristic. For example,because weight and height have differentdistributions (variances), -2 Z-scores onthe weight-for-age distribution is about80% of the median, and -2 Z-scores on theheight-for-age distribution is about 90% ofthe median. Further, the proportion of thepopulation identified by a particularpercentage of the median varies at differentages on the same index.

The second major advantage of using Z-scores is that useful summary statistics can

be calculated from them. The approachallows the mean and standard deviation tobe calculated for the Z-scores for a group ofchildren. This can not be done withpercentages of the median or centiles.Thus, the Z-score application is consideredthe simplest way of describing thereference population and makingcomparisons to it. It is the statisticrecommended for use when reportingresults of nutritional assessments.Examples of Z-score calculations arepresented in Appendix 1.

The Z-score or standard deviationunit (SD) is defined as the differencebetween the value for an individual and themedian value of the reference populationfor the same age or height, divided by thestandard deviation of the referencepopulation. This can be written in equationform as:

6.4. Percentage of the Median and Percentiles

The percentage of the median is definedas the ratio of a measured or observedvalue in the individual to the medianvalue of the reference data for the same

age or height for the specific sex,expressed as a percentage. This can bewritten in equation form as:

(observed value) - (median reference value)

standard deviation of reference populationZ-score (or SD-score) =

observed value

median value of reference populationPercent of median =

100


6.

41

The median is the value at exactly themid-point between the largest andsmallest. If a child’s measurements areexactly the same as the median of thereference population we say that they are“100% of the median”. Examples ofcalculations for percent of median can befound in Appendix 1.

The percentile is the rank position ofan individual on a given referencedistribution, stated in terms of whatpercentage of the group the individualequals or exceeds. Percentiles will not bepresented in this guide. The distributionof Z-scores follows a normal (bell-shapedor Gaussian) distribution. The commonlyused -3, -2, and -1 Z-scores are, respect-

ively the 0.13th, 2.28th, and 15.8thpercentiles. The percentiles can bethought of as the percentage of children inthe reference population below theequivalent cut-off. Approximately 0.13percent of children would be expected tobe below -3 Z-score in a normallydistributed population.

Z-score Percentile

-3 0.13

-2 2.28

-1 15.8

6.5. Cut-offs

The use of a cut-off enables the differentindividual measurements to be convertedinto prevalence statistics. Cut-offs are alsoused for identifying those childrensuffering from or at a higher risk ofadverse outcomes. The children screenedunder such circumstances may beidentified as eligible for special care.

The most commonly-used cut-off withZ-scores is -2 standard deviations,irrespective of the indicator used. Thismeans children with a Z-score forunderweight, stunting or wasting, below -2 SD are considered moderately orseverely malnourished. For example, achild with a Z-score for height-for-age of -2.56 is considered stunted, whereas achild with a Z-score of -1.78 is notclassified as stunted.

In the reference population, bydefinition, 2.28% of the children would bebelow -2 SD and 0.13% would be below -3SD (a cut-off reflective of a severecondition). In some cases, the cut-off fordefining malnutrition used is -1 SD (e.g. inLatin America). In the reference orhealthy population, 15.8% would be belowa cut-off of -1 SD. The use of -1 SD isgenerally discouraged as a cut-off due tothe large percentage of healthy childrennormally falling below this cut-off. Forexample, the 1995 DHS survey using a –2SD cut-off for stunting in Uganda found a

36% prevalence of stunting in under-threeyear olds. This level of stunting is about 16times the level of the reference population.

A comparison of cutoffs for percent ofmedian and Z-scores illustrates thefollowing90% = -1 Z-score80% = -2 Z-score70% = -3 Z-score (approx.)60% = -4 Z-score (approx.)

Cut-off Points for MUAC for the 6 - 59Month Age GroupMUAC cut-offs are somewhat arbitrarydue to its lack of precision as a measure ofmalnutrition. A cut-off of 11.0cm can beused for screening severely malnourishedchildren. Those children with MUACbelow 13.5cm should be selected to havetheir weight-for-height measured.

Global Malnutrition is a term generallyused in emergency settings. The globalmalnutrition rate refers to the percent ofchildren with weight-for-height below -2 Z-scores and/or edema. This refers to allmoderate and severe malnutritioncombined. The combination of a lowweight-for-height and any child with edemacontributes to those children counted as inthe global malnutrition statistic.

C O M PA R I S O N O F A N T H RO P O M E T R I C DATA TO R E F E R E N C E S TA N DA R D S PA RT 6 .


42

6.

Malnutrition Classification SystemsThe cut-off points for differentmalnutrition classification systems arelisted below. The most widely used systemis WHO classification (Z-scores). TheRoad-to-Health (RTH) system is typicallyseen in clinic-based growth-monitoringsystems. The Gomez system was widelyused in the 1960s and 1970s, but is onlyused in a few countries now. An analysisof prevalence elicits different results fromdifferent systems. These results would notbe directly comparable. The difference is

especially broad at the severe malnutritioncut-off between the WHO method (Z-scores) and percent of median methods. At60% of the median the closestcorresponding Z-score is –4. The WHOmethod is recommended for analysis andpresentation of data (see Part 6.2).

Mild, moderate and severe are differentin each of the classification systems listedbelow. It is important to use the samesystem to analyze and present data. TheRTH and Gomez classification systemstypically use weight-for-age.

System Cut-off Malnutrition classification

WHO < -1 Z-score mild

< -2 Z-score moderate

< -3 Z-score severe

RTH ≥ 80% of median normal

60% - < 80% of median mild-to-moderate

< 60% of median severe

Gomez ≥ 90% of median normal

75% - < 90% of median mild

60% - < 75% of median moderate

≤ 60% of median severe


7.

43

7.Data Analysis

PA

RT

Computer software can be used to makecomparisons to the reference standards. TheCenters for Disease Control and Prevention(CDC) has developed a free softwarepackage called Epi Info that can handle all ofthese anthropometric calculations.Cooperating Sponsors are stronglyencouraged to use available software toanalyze nutrition data. Not only will thesoftware enable raw anthropometric surveydata to be transformed into the indices andscores described in Part 6, the software willflag outliers which are usually the result ofincorrect measurements, coding errors orincorrect ages. Once the anthropometricindices have been calculated, they can bepresented in simple tables using specifiedcut-offs and age categories consistent withthe Title II Generic Indicator list.

When using computer software foranthropometry, there are three separateprocedures that should be performed.First, the raw measurement data should beentered into the computer. Second, theprogram should combine the raw data onthe variables (age, sex, length, weight) tocompute a nutritional status index such asweight-for-age, height-for-age or weight-for-height. Third, the program shouldtransform these data into Z-scores so thatthe prevalence of nutritional conditionssuch as being underweight and stunted canbe calculated.

Another software called ANTHROanalyzes anthropometric data and can bedownloaded from the WHO Global Database

on Child Growth and Malnutrition(www.who.int/nutgrowthdb).

There is another Windows basedsoftware available from: www.nutrisurvey.de.This software program was designedspecifically for nutrition surveys by the WorkGroup on International Nutrition of theUniversity of Hohenheim/Stuttgart incooperation with the German Agency forTechnical Cooperation (GTZ). Thesoftware is based on the Guidelines forNutrition Baseline Surveys in Communitiespublished by GTZ. The purpose of theprogram is to integrate all steps of aNutrition Baseline Survey into a singleprogram. The program contains astandard Nutrition Baseline questionnairewhich can be easily customized for thespecific site, a function for printing out thequestionnaire, a data entry unit whichcontrols the data being entered, a speciallyadapted plausibility check, a reportfunction and a graphics section. The reportfunction produces the full set of descriptivestatistics of a baseline survey. The graphicssection contains standard graphs andadditional graphics for the anthropometricindices with comparison to the NCHSstandard. The anthropometric indices (Z-scores of height-for-age, weight-for-height,weight-for-age) and the prevalence ofstunting, wasting, underweight andoverweight of children are calculatedautomatically. For further statisticalevaluation, the data can be exported toSPSS or other statistical programs.


44

7.

7.1. Sources of Epi Info Software

Public domain sources of Epi Infosoftware and supporting materials

Epi Info is available from the Centersfor Disease Control and Prevention (CDC),1600 Clifton Road, Atlanta, GA 30333, USA

or downloaded from:www.cdc.gov/epiinfo/

The Epi Info 2000 package comes witha manual and tutorials to help the user tobecome familiar with data analysis usingEpi Info.

7.2. Recommendation for Analysis and Presentation of Height Data

For evaluation purposes, the presentation ofstunting data for children less than 24months is useful. An intervention amongchildren under 24 months is likely to bemore effective than among children 24-59months. This is because 1) the deter-minants of stunting in the older childrenare more varied, and 2) stunting in olderchildren may reflect historical nutritional orhealth stress and be ‘permanent,’ i.e. notresponsive to any intervention. A furtherconsideration in the presentation of data onstunting by age groups is the change inmeasurement technique at 24 months ofage. Data on length for children less than24 months should not be grouped together

with data on height for children morethan 24 months of age.

Prevalence - For use in Title II programs,the prevalence of nutritional statusconditions can be calculated using cut-offpoints for height-for-age and weight-for-age. The cut-off points can be set using Z-scores, percentiles or percentage of themedian. For Title II programs, a cut-off of -2 Z-score is recommended and resultsshould be presented for both males andfemales. An example of a prevalence tablefor low height-for-age as established by -2SD for groups of children aged 6 - 59.99months is given below:

Table 7.1. Prevalence of low height-for-age (stunting) in a sample of 97 children, by sex and age group

Age Group(months)< 6

6-11.99

12-23.99

24-35.99

36-47.99

48-59.99

Total

SexBoysGirlsCombinedBoysGirlsCombinedBoysGirlsCombinedBoysGirlsCombinedBoysGirlsCombinedBoysGirlsCombinedBoysGirlsCombined

Number belowcut-off (-2 SD)0000001230446396511131427

Number in agegroup6410639610167132010102015722504797

Percentage belowcut-off0.00.00.00.00.00.016.720.018.80.030.820.060.030.045.040.071.450.026.029.827.8

Overall, this table shows that 26.0 percent of the boys had a low height-for-age or were stunted, while 29.8 percent of girls were stunted. Interpretations can also be made about the various age groups or with boys and girls grouped together.


DATA A N A LYS I S PA RT 7 .

7.

45

7.3. Examples of Data Analysis

This section covers an analysis of datapertaining to maternal and child nutritionprograms using food aimed at vulnerablegroups in two populations (countries A andB), simulating situations often encounteredin Title II operations. These examplesillustrate an analysis that would allow one toreach, at a minimum cost, statistically validconclusions concerning the nutritionalimpact of a supplementary feeding program.

The analysis compares the “beforeintervention” and “after intervention” dataaccording to the following plan going fromthe general to the specific.

For all ages and sexes combined:

• all ages, both sexes comparison ofchange in the indicators;

• calculation of general prevalence ofmalnutrition (stunting and underweight);

• calculation of changes in prevalence;

Examination of data separated by sex(ages combined):

• calculation of general prevalence ofmalnutrition for each sex;

Example 1

In 1995, a US-supported Cooperating Sponsor in cooperation with the Government of Country A,introduced a community-based health and nutrition program with a supplementary feeding schemeaimed at vulnerable groups. The scheme covered 6 of the 14 administrative districts of the country.Food supplements were distributed through community centers on a year-round basis to infants oversix months of age and children up to the age of 3 years and to pregnant women and lactating mothersover an 18-month period (last 6 months of pregnancy and the first 12 months postpartum). In all,86,000 individuals (21,000 women and 65,000 children) were covered by the program.

In agreement with the USAID Mission and Food for Peace (FFP), the Cooperating Sponsor decided toevaluate the nutritional changes at three intervals (baseline, mid-term and final year). The firstcollection of data began in the second year following the initiation of operations. By adopting a pre-post or reflexive design, data on sex, age, weight and height were collected from three representativecross-sectional sample surveys of the infant and child beneficiaries in 1996, 1998 and 2000.

Ages of the sample children ranged from 3 months to 5 years. Due to some problems in ageestimations and incomplete data, sample sizes of the children varied between 3700 and 2500. No datawere collected from the 8 districts in which the program had not been implemented.The populationof the 8 districts was not comparable, from the nutritional and socioeconomic standpoints, with thepopulation of the 6 districts covered by the program.

• calculation of changes in prevalence;

Examination of data by age categories(sexes combined):

• calculation of prevalence of mal-nutrition by age category;

• calculation of changes in prevalence.

With reference to this plan of analysis, thefollowing assumptions should be noted:

• the final year evaluation measure-ments were taken after an intervalsufficiently long for the program to haveproduced a nutritional impact (e.g. fiveyears);

• data were collected from a cross-sectional sample representative of the pro-gram population; and

• baseline data had been collected.

The first example is a comparison ofpercentage or prevalence changes.Cooperating Sponsors are encouraged tocompare changes in the mean Z-scoresfor statistical and epidemiological rigor.


46

7.

Calculation of Nutrition LevelsThe measurement of nutritional impactwas based on a comparison of datacollected in 1996 and 2000 so that thefigures would include the largest possibleproportion of children who hadparticipated in the program for one year ormore. It was felt that this time intervalwas necessary to allow for any anticipatedimpact to manifest itself. The collecteddata are summarized in Table 7.5.

The results for weight-for-height arepresented to illustrate that this indicatoris inappropriate to evaluate the programsince it reflects short-term changes. Thetables also include age ranges that willnot be used in the final presentation but

Cooperating Sponsors are encouraged toexamine the results for different agegroups to better understand how theindicator responds.

Table 7.2. Distribution of nutritional indicators (Z-scores) at baseline (all ages)

Height-for-ageHAZ

2695

1294

48.0

No. examined

No. below -2 S.D.

% below -2 S.D.

Weight-for-ageWAZ

2695

916

34.0

Height-for-heightWHZ

2695

110

4.1

Table 7.3. Prevalence of low levels of nutritional indicators by sex at baseline (all ages)

HAZ < -2

49.5

47.3

Girls

Boys

WAZ < -2

32.4

35.6

WHZ < -2

3.6

4.6

Table 7.4. Prevalence of stunting or low levels of height-for-age by agecategories at baseline

3-11 months

34.5

560

HAZ < -2

No. examined

12-23 months

42.4

523

36-59 months

52.7

1078

24-35 months

48.5

534

Table 7.5. Prevalence of low levels of nutritional indicators by sex and stage ofintervention (for specific age categories)

HAZ < -224-59 months

Baseline

49.5

47.3

48.0

Girls

Boys

Sexes combined

35.5

37.5

36.6

32.4

35.6

34.0

23.1

26.5

25.0

3.6

4.6

4.1

3.2

4.0

3.8

Final year

WAZ < -23-35 months

Baseline Final year Baseline Final year

WHZ < -23-35 months


7.

47

Tables 7.4 and 7.5 show that over the fouryears of the intervention, stunting in girlsand boys was substantially reduced alongwith underweight. For sexes combined,the reduction was 11.4 and 9.0 percentagepoints for stunting and underweight,respectively. The reduction for stuntingwas more dramatic while wasting (WHZ)was virtually unaffected. Wasting shouldnot be used for evaluation purposes as it isa relatively rare event and very susceptibleto seasonal influences.

Note the different age groupings forstunting and wasting in Table 7.5. Some ofthe children measured for stunting wereolder than the children in the intervention.The reason for selecting the 24-59 monthage group for evaluation was to capturethe cumulative and lagged effect that thenutrition project would have on stunting.

There is another reason for thedifferent age groupings. It is notrecommended to aggregate data forchildren under 24 months with those over24 months (see Part 7.2). Also, the Title IIGeneric Indicators recommend eitherstunting or underweight indices butrequire specific age groupings forunderweight.

Comparison of Mean Z-scoresThe alternative and preferred approach toevaluating the change in a percentage fora nutritional index is to compare mean Z-score change over the life of the program.Just as in the above calculation of changein prevalence of an index, the data are

analyzed at baseline with the mean andstandard deviation calculated andcompared with the same project area inthe final year of the program. The meanZ-score comparison has the advantage ofdescribing the entire population directly,without resorting to a subset of indi-viduals below a set cut-off. Comparingmeans over prevalences is desired asmany of the Title II interventions targetwhole communities not just the severelymalnourished.

A community health and nutritionintervention would expect all children tobenefit whereas a targeted feedingprogram for the severely malnourishedwould only benefit these children. Using a-2 SD cut-off and presenting a prevalencechange would show a change in theprevalence of those below the cut-off.Therapeutic feeding programs wouldfocus on changes in nutritional statusamong the severely malnourished butcommunity based programs target allchildren and their caregivers. Apresentation of the mean would reflect allthe children and comparing means wouldreflect the community shift or improve-ment. The statistical comparison of meanZ-scores over time using the Student’s T-test for example, is a more powerfulstatistical test than comparing prevalencesusing the Chi-square statistical test. Usingthe same example from above, Table 7.6presents the results of the mean Z-scoresfor height-for-age and weight-for-age.

Table 7.6. Mean Z-scores for stunting (HAZ) and underweight (WAZ) by sex and stage ofintervention (for specific age groupings)

Girls

Boys

Sexes combined

-1.98

-1.86

-1.93

-1.54

-1.65

-1.58

WAZ 3-35 months

Baseline Final year

-2.53

-2.48

-2.50

-2.33

-2.45

-2.36

HAZ 24-59 months

Baseline Final year

The evaluation recommendation for TitleII programs is to use a comparison ofmean Z-scores for statistical testing butthe results should be presented with both

change in mean Z-score and change inprevalence as the latter is more easilyunderstood by a general audience.

DATA A N A LYS I S PA RT 7 .


48

7.

Example 2

A Cooperating Sponsor conducted a baseline and final-year survey collecting height and weight data on 24-59

month old children. At both times the sample was randomly drawn from the target communities and did not include

the same children in both surveys. The pre-post design enabled a comparison of change in nutritional status (as

reflected by height-for-age) in the target communities.

Baseline at time zero:

(t0): Mean Z-score = -2.05 (sd=1.26); Prevalence (-2SD cut-off) = 40%

Sample size = 940

Final year at time five:

(t5): Mean Z-score = -1.20 (sd=1.15); Prevalence (-2SD cut-off) = 23%

Sample size = 1056

Using statistical software to conduct the t-test (e.g. SPSS, or STATA; note that Epi Info 2000 or Epi Info 6 does not

have this test), the testing of the significance of the change in the sample means is straightforward.

t = (Mean t0 - Mean t5) / sq. root (Variance t0/n0 + variance t5/n5)

where n0 and n5 is sample size at baseline and final year.

t = (-2.05 - (-1.20)) / [sq. root ((1.26x1.26)/940)) + ((1.15x1.15)/1056))]

t = (-0.85) /[sq. root (0.0016889) + (0.0012523)]

t = (-0.85) /[ 0.054233]

t = 15.67

This change in mean Z-scores is highly significant (i.e. well above the critical t-value of 1.96 at 0.05 level).

7.4.Additional Data Analysis Information

In cooperation with Food Aid Management,FANTA developed a workshop for thetraining of program managers in thecalculation and analysis of basicanthropometric data using Epi Info andSPSS. The materials from the Data AnalysisWorkshop are available at:www.fantaproject.org/focus/index underMonitoring and Evaluation.

Step by step instructions on data analysiscan be found on the Practical Analysis ofNutritional Data (PANDA) website:www.tulane.edu/~panda2/. PANDA wasdesigned for data analysis instruction usingSPSS software.

There are many statistical analysis softwarepackages available, other than Epi Info. Thepackages listed below are only a few of thecommonly used statistical analysis softwareprograms. These packages vary incapability and cost. Information oncapability and ordering the packages can befound on their web sites.SPSS (Statistical Package for SocialSciences) www.spss.com SAS (Statistical Analysis System)www.sas.comSTATA (Statistics/Data Analysis)www.stata.comSUDAAN (Software for the StatisticalAnalysis of Correlated Data)www.rti.org/patents/sudaan/sudaan.html


8.

49

8.AnnualMonitoringIndicators

PA

RT

8.1. Introduction

Information from regular reporting ofactivities from growth monitoring andpromotion programs (GMP) will enhanceprogram management and can providevaluable insights into the interpretation ofanthropometric indicators of impact thatare required by USAID. This sectiondescribes how annual monitoringindicators that are based upon data frommonitoring may be collected and reportedin a standard format. This will help makethe indicators more useful formanagement of programs at all levels

within countries and for reporting toUSAID. Reporting information to USAIDon annual monitoring indicators isrecommended but not required.

The recommended annual monitoringindicators for maternal and child healthand nutrition programs are:

• Percent of eligible children in growthmonitoring/promotion (MP);

• Percent of children in GMP programgaining weight in past 3 months (genderspecific).

8.2. Routine Data Collection

These guidelines are designed to avoid anyunnecessary burden of reporting on staff atall levels in program implementation andso information for the monitoringindicators should come from routinelycollected data rather than special surveys.At this time, there is not a standard formatfor collecting data from GMP activity and,the indicators that are available will varywidely, among Cooperating Sponsors andcountries, and in some cases within thesame Cooperating Sponsor. It isappropriate to report whatever data arereadily accessible and avoid investingscarce resources in attempts to generate orretrieve data that are not readily accessible.

Data needed for the first index, GMPattendance rate:

1. The denominator - the number ofeligible children in the population, bygender and age (usually the most reliableestimates will be those the project hasgathered as part of its baseline survey, orother data collection activities);2. The numerator - the number ofchildren in growth monitoring/promotion.The classification of “in growthmonitoring/promotion” is meant to reflectthe total number of children whose weighthas been monitored in clinic- and/orvillage-based activities. This number maybe the total number of children weighed inthe ‘round’ immediately before thereporting period, or it may be an estimateof the ‘usual’ total number of childrenattending weighing activities over thereporting period. How the number is


50

8.

derived is not necessarily important, but itis most important to report clearly howthe estimate was made. Becauseattendance at GMP programs varieswidely with age, this number will be moremeaningful if it is age specific - e.g., <12months, 12-<24 months, 24-<60 months(ages reported will be influenced by thetarget age group of the project).

The definition of the numerator of the firstindex is central to the definition of thesecond index.

Data needed for the second index, GMPincreased weight rate among enrolledchildren:

1. The denominator -- this will be thesame as the numerator from the firstindex, i.e. the total number of childrenattending growth monitoring/promotion;and2. The numerator -- the number ofchildren in growth monitoring/promotionwho gained weight in the last 3 months.Two elements should be considered butthe direction of weight change is moreimportant than the second concerning the3-month time frame.

The reports of many weighing activities,both clinic- and village-based, include asummary that presents a) the total numberof children attending, b) the number ofchildren who gained weight, and c) thenumber of children who did not gainweight. All children weighed should beclassified as either gaining weight or notgaining weight - usually over a three monthperiod. Calculate the index by dividing thenumber of children who gained weight bythe total number of children attending,then multiply this by 100. This index

should be considered available only inthose programs in which the appropriatesummary numbers are reported. In somesituations a time frame other than 3months is the only available information.

Collecting accurate data from weighingis difficult. To ensure the quality of thedata, health workers should be properlytrained to make accurate measurements ofwhether or not a child is gaining weight.Data from assessments are central to theusefulness of growth monitoring andpromotion programs.

Annual benchmarks are notappropriate for these indicators becausetheir interpretation is specific to thecontext of the particular program and itsactivities. For example, when a programextends activities into areas of highestneed (perhaps because of remoteness orfood insecurity) the overall percent ofchildren in growth monitoring/promotionand the percent of children gaining weightare likely to decrease. Clearly this resultwill not be interpreted as reflecting poorlyon program implementation. Thisexample demonstrates that the monitoringindicators proposed here could not beinterpreted independent of the context ofthe activities of a program. Theseindicators are not useful for summativeevaluation and are not intended to be usedfor this purpose.

Substantial resources should not beinvested to gather data on annualmonitoring indicators reported byCooperating Sponsors. However, it isrecommended that Cooperating Sponsorsconsider the advantages of including themonitoring indicators in their annualreports and to modify training,implementation and information manage-ment systems to incorporate theseindicators in the future.

8.3. Data on Growth Monitoring and Promotion

Relevant items could be added toquestionnaires at baseline and end-line todescribe changes over the time period of anintervention (provides an assessment ofimpact of the intervention on service deliverythat might help support or explain findings

from anthropometric status of children).Resources on GMP are available in aDiarrhea Dialogue supplement on GrowthMonitoring Health Basics: Issue no. 24 -March 1986 at: www.nand.org/dd/html/su24.htm.


9.

51

9.References

PA

RT

Adair, L., Polhamus, B., Testing Indicators for Use in MonitoringInterventions to Improve Women’sNutritional Status. MEASURE EvaluationProject, The United Nations UniversityPress, May 1998.

Beaton, G., Kelly, A., Kevany, J.,Martorell, R., and Mason, J.Appropriate Uses of AnthropometricIndices in Children. Geneva: UNACC/SCN, 1990

Bender, W., and Remancus, S. The Anthropometry Resource Centerfunded by the FAO/SADC projectGCP/RAF/284/NET, Development of aRegional Food Security and NutritionInformation System.

CDC. ANTHRO software analyzesanthropometric data can be downloadedat www.cdc.gov/epiinfo/ and from theWHO Global Database on Child Growthand Malnutrition atwww.who.int/nutgrowthdb

Collins, S., Duffield, A., and Myatt, M.Assessment of nutritional status inemergency-affected populations: Adults.ACC/SCN, Geneva, July, 2000

Dean, A. G., Epi Info Version 6: A WordProcessing Database and Statistics Systemfor Epidemiology on Microcomputers,Atlanta, Georgia, 1994.

Food and Nutrition BulletinSurveillance for actions towards betternutrition, Volume 16, Number 2, TheUnited Nations University Press, June1995. www.unu.edu/unupress/food/8F162e/8F162E00.htm#Contents

FAO (Food and AgricultureOrganization), Conducting small-scalenutrition surveys: A field manual, No. 5,FAO, Rome, 1990.

Hennekens, C., Burring, J.Epidemiology in Medicine. Little, Brownand Company. 1987.

Ismail, S., Manandhar, M., BetterNutrition for Older People, Assessment andAction. HelpAge International andLondon School of Hygiene & TropicalMedicine, London, UK, 1999.

Médecins sans Frontèries (MSF),Nutrition Guidelines. Coordinated by MBoerlart, A Davis, B Lelin, M J Michelet,K Ritmeijer, Svan der Kam and F Vautier.MSF Holland, 1995.

Médecins sans Frontèries (MSF),Refugee Health, An Approach toEmergency Situations, 1997.

Norusis, M., Marketing Department,SPSS Inc. 1990.


52

9.

Pellitier, D., Frongillo Jr., E, Schroeder,D., Habicht, J. A Methodology forEstimating the Contribution ofMalnutrition to Child Mortality inDeveloping Countries. The Journal ofNutrition. October 1994, Supplement.

SCF Drought Relief in Ethiopia. Compiledby J Appleton with the SCF EthiopianTeam, London. 1987.

Starr, C., McMillan, B., Human Biology,Third Edition. Brooks/Cole PublishingCompany, 1999.

United Nations Department ofTechnical Co-operation forDevelopment and Statistical Office,How to Weigh and Measure Children:Assessing the Nutritional Status of YoungChildren in Household Surveys, UN, NewYork, 1986.

UNICEF, 1995. Monitoring progress towardthe goals of the World Summit for Children:A practical handbook for multiple-indicator surveys, UNICEF, New York.

UNICEF equipment specifications can befound at: www.supply.unicef.dk/catalogue/index.htmunder 03 NUTRITION

USAID/BHR, Commodities ReferenceGuide (CRG), 2000.www.usaid.gov/hum_response/crg.

USAID/BHR/FFP. Title II Guidelines forDevelopment Programs. January 2000.www.usaid.gov/hum_response/ffp/dappaa.htm

USAID/CDIE. Performance Monitoringand Evaluation Tips. 1996.www.usaid.gov/pubs/usaid_eval/#02.

Waterlow JC. Classification anddefinition of Protein Calorie malnutrition.British Medical Journal 3:566-569, 1972

Waterlow JC, Buzina R, Keller W, LaneJM, Nichaman MZ, Tanner JM.The presentation and use of height andweight data for comparing the nutritionalstatus of groups of children under the ageof 10 years. Bulletin of the World HealthOrganization 55(4):489-498., 1977

WHO (World Health Organization)Measuring Change in Nutritional Status.Guidelines for Assessing the NutritionalImpact of Supplementary FeedingPrograms for Vulnerable Groups. WorldHealth Organization, Geneva, 1983.

WHO (World Health Organization)Management of Severe Malnutrition: AManual for Physicians and Other SeniorHealth Workers. World HealthOrganization, Geneva, 1999.


9.

53

WHO (World Health Organization)The Management of Nutrition in MajorEmergencies. World Health Organization,Geneva, 2000.

WHO (World Health Organization)Field Guide on rapid nutritionalassessment in emergencies. WHO RegionalOffice for the Eastern Mediterranean,World Health Organization, Geneva,1995.

WHO (World Health Organization)Physical Status: The use and interpre-tation of anthropometry. Report of aWHO Expert Committee. World HealthOrganization, Geneva, 1995.

Woodruff, B.A. and Duffield, A.Assessment of nutritional status inemergency-affected populations:Adolescents. ACC/SCN, Geneva, July, 2000

World Food Programme (WFP).Supplementary Feeding for Mothers andChildren: Operational Guidelines also seeUNHCR/WFP Guidelines For SelectiveFeeding Programmes In EmergencySituations, February 1999www.wfp.org/OP/guide/PolGuideSelect.html.Useful Websites

ACC/SCN (Nutrition Sub-committee ofthe UN): acc.unsystem.org/scn/

Anthropometric Desk Reference:www.odc.com/anthro/

Center for Disease Control (CDC)Growth Charts:www.cdc.gov/growthcharts/

Child Survival Technical SupportProject (CSTS) for KPC Material andother useful information:www.childsurvival.com

CORE: USAID/PVO Core Group:www.coregroup.org/

Demographic Household Survey (DHS)Macro Inc. www.measuredhs.com/

Food and Agriculture Organization(FAO) Nutrition Division:www.fao.org/WAICENT/FAOINFO/ECONOMIC/ESN/NUTRI.HTM

Food Aid Management (FAM) providesUSAID documents related to Title IIprograms for PVOs: www.foodaid.org

Food and Nutrition TechnicalAssistance Project: www.fantaproject.org

International Life Sciences Institute(ILSI): www.ilsi.org/

London School of Hygiene and TropicalMedicine, Public Health Nutrition Unitwww.lshtm.ac.uk/eps/phnu/phnintro.htm

R E F E R E N C E S PA RT 9 .


54

9.

LSMS World Bank Surveys:www.worldbank.org/html/prdph/lsms/index.htm

MEASURE I (DHS):www.measuredhs.com/

MEASURE II Evaluation USAIDProject:www.cpc.unc.edu/measure/home.html

Nutrition Surveys and Assessment:www.nutrisurvey.de/

Practical Analysis of Nutrition Data(PANDA) Tulane University:www.tulane.edu/~panda2/

USAID Commodity Reference Guide,Part Two:www.usaid.gov/hum_response/crg/

WHO Global Database on ChildGrowth and Malnutrition:www.who.int/nutgrowthdb/


9.

55

Acute undernutrition - Serious and usuallycurrent periods of inadequate food intake.

Age chart - A chart that can be usedquickly to determine a child’s current ageby the month and year the child was born;a tool used to determine if reported birthdates match the age of a child given byparents or estimated from the child’sappearance. (FAO, 1990)

Anthropometry - The study andtechnique of taking body measurements,especially for use on a comparison orclassification basis.

Arm circumference - A measurementdone on the mid-upper arm; ameasurement used to assess total bodymuscle mass and in some circumstances,protein-energy malnutrition.

Asymmetrical - Being lopsided; nothaving a equal correspondence of formand arrangement of parts on oppositesides of a boundary; an asymmetricaldistribution would not have two equalhalves on each side of the median.

Bar chart - A chart in which the lengthof the bars depends on the number ofcases in that category and the number ofbars depends on the number ofcategories.

Batch processing - The processing of datafor a large group of people at one time.

Bias - A consistent, repeated difference ofthe sample from the population, in thesame direction; sample values that do notcenter on the population values but arealways off in one direction.

Body Mass Index (BMI) - Also known as“Quetelet’s index”. An index that uses thevariables weight and height to measurebody fat stores, usually in adults ratherthan children (weight in kilogramsdivided by the square of height in meters).

Case-control study - A study in whichsubjects are selected on the basis ofwhether they are (cases) or are not(controls) receiving benefits of a healthand, or nutrition program.

Chi-Squared test - The Chi-Squared Testof Association looks at the statisticalsignificance of an association between acategorical outcome (such as wasted ornot wasted) and a categorical determiningvariable (such as diarrhea in the last twoweeks, no diarrhea).

Chronic undernutrition - Extendedperiods of inadequate food intake.

Circumference measuring tape(circumference insertion tape) - A toolused to assess arm circumference; a


Glossary


56

9.

plastic, non-stretchable tape that is pulledtaut around the mid-point of the upperarm to measure circumference of the arm.

Classification system - A system thatestablishes cut-off points usingpercentiles, percentages of the median orstandard deviations and identifiesdifferent levels of nutritional risk.

Cluster sample - The selection of groupsthat are geographically close to oneanother for a sample; usually used ininstances when lists of households orindividuals are not readily available.

Cohort studies - A study which focuseson the same group of people, but usesdifferent individuals over time; a studythat uses the same specific populationeach time but uses different samples.

Confidence interval - An interval that hasa specified probability of covering the true population value of a variable orcondition.Cross-section plus over-sample survey -A survey in which data is collected from arandom sample and then additional datais collected so that an in-depth view canbe gained of a certain group or problem.

Cut-off point - Predetermined risk levelsused to differentiate betweenmalnourished and adequately nourishedsegments of a population.

Design effect - The loss of samplingefficiency resulting from the use of clustersampling instead of random sampling (adesign effect of 2.0 is commonly used foranthropometric and immunization surveys).

Distribution - A display that shows thenumber of observations (or measure-ments) and how often they occur.

Edema - The presence of excessiveamounts of fluid in the intercellular tissue.It is the key clinical sign of a severe formof protein energy malnutrition.

Epidemiology - The science of theoccurrence and determinants of disease ina population. Epi Info software - A series of micro-computer programs produced by the CDCand WHO, for handling epidemiologicaldata in questionnaire format and for organ-izing study designs and results into text andtables that may form part of written reports.

External validity - Being able togeneralize conclusions drawn from asample or sub-set to a wider population.

Gomez classification system - Aclassification system that uses percentageof the median weight-for-age to identifychildren as being normal or having mildmalnutrition, moderate malnutrition orsevere malnutrition.


9.

57

Graph - A drawing that shows therelationship between two sets of numbersas a set of points having coordinatesdetermined by their relationship; a displayof numerical relationships.

Growth chart - A graph that is usuallyused to record a child’s weight-for-age inmonths; a chart typically used by mothersand health workers to determine if a child is experiencing a normal gain in weight.

Growth faltering - A condition identifiedby emphasizing the direction of growth obtained in serial recordings, rather thanactual weight-for-age itself; signified by nochange or an actual decrease inmeasurements. (ACC/SCN, 1990)

Growth monitoring and promotion -The practice of following changes in achild’s physical development, by regularmeasurement of weight and sometimes oflength with accompanying information toguide the care givers’ nutritional andrelated care.

Histogram - A display that shows thenumber of observations and how oftenthey occur, usually through the use ofvertical bars and a horizontal base that ismarked off in equal units.

Household - One person who lives aloneor a group of persons, related orunrelated, who share food or make

common provisions for food and possiblyother essentials for living; the smallestand most common unit of production,consumption and organization insocieties.

Index - An index is usually made up oftwo or more unrelated variables that are used together to measure an underlying characteristic.

Indicator - A measure used at thepopulation level to describe the proportionof a group below a cut-off point; example:30 percent of the region’s children arebelow -2 SD for height-for-age.

Intrahousehold distribution - Thedistribution of food within a household;the act of determining what proportion ofthe total household food supply eachmember of the household receives.

Length-for-age - An index of past orchronic nutritional status; an indexwhich assesses the prevalence ofstunting.Local events calendar - A calendar thatreflects important local events andseasons that might help a parent pinpointthe birth date of their child.

Longitudinal survey - A survey whichfollows people over time, to capture dataon an evolving situation or problem.Different types of longitudinal surveys



58

9.

include: cohort studies, trend studies andpanel studies.

Malnutrition - A nutritional disorder orcondition resulting from faulty orinadequate nutrition.

Mean - The average value for a set of data;a measure of central location obtained byadding all the data items and dividing bythe number of items.

Measurement error - The error that canresult in a survey from incorrect(anthropometric) measurements beingtaken.

Median - A measure of central locationfor a set of data; the value that falls in themiddle of a set of data when all the valuesare ordered from lowest to highest.

Morbidity - A condition resulting from orpertaining to disease; illness.

Mortality rate - Death rate; frequency ofnumber of deaths in proportion to apopulation in a given period of time; death.

NCHS reference standards - Growthpercentiles developed by the NationalCenter for Health Statistics in the US thatprovide standards for weight-for-age, length-for-age and weight-for-length.

Normal distribution - A normaldistribution takes a bell-shape and has thefollowing characteristics: the highest pointoccurs at the mean; it is symmetric; thestandard deviation determines the widthof the distribution; and it can be describedwith only two numbers: the mean and thestandard deviation.

Numeric value - A value expressed as anumber or numeral.

Nutritional surveillance - A system ofdata collection and application; systemsthat are based on routinely compiled dataand that monitor changes in variables overtime, give warning of impending crisis ormonitor the effectiveness/ineffectivenessof existing programs and policies; thecontinuous monitoring of the nutritionalstatus of a specific group.

One-tailed Test - A statistical test to detecta difference in means between twopopulations in a specified direction (i.e. todetect improved nutritional status).

One time assessment - The practice ofassessing nutritional status through theuse of measurements taken on oneoccasion, usually used to screenparticipants for immediate interventions.

Panel studies - A type of longitudinal surveythat studies the same people over time.


9.

59

Percentiles - A number that correspondsto one of 100 equal divisions in a range ofvalues; a measure of relative location; avalue such that at least p% of the items inthe data set are less then or equal to itsvalue and at least (100 - p)% of the itemsare greater than or equal to it; example:the 60th percentile means that 60% ofvalues in the data set are less than or equalto it and (100 - 60) 40% are greater than orequal to it.

Percentage of the median - A fraction orratio based on a total of 100, where themedian value of the data set equals 100; avalue that equals a proportion or part of adistribution where the median represents100 percent.

Population - The entire group of peoplethat is the focus of the study (everyone inthe country, or those in a particularlocation, or a special ethnic, economic orage group).

Prevalence - The proportion of thepopulation that has a condition of interest(i.e. wasting) at a specific point in time; ameasure of a condition that is independentof the size of the population; a value that isalways between 0 and 1.

Protein-energy malnutrition - Under-nutrition that results in an individual notreceiving adequate protein or calories fornormal growth, body maintenance, and the

energy necessary for ordinary humanactivities.

References or reference standards -Measurement data collected on represent-ative, healthy populations throughstandardized methods; a data set thatallows comparisons to be made betweenits values and individuals or populationsbeing measured.

Reflexive study design – One group pre-/post-test. A study in which one populationgroup is studied on two different occasionsto compare changes over time.

Risk - The possibility of suffering harm;danger; “a continuous variable relating tothe likelihood that a defined undesirableoutcome will occur”.

Sample - A part or subset of thepopulation used to supply informationabout the whole population.

Sample size - The number of householdsor persons selected to be included in asample or survey.

Sampling - The technique of selecting arepresentative part of the population forthe purpose of determining characteristicsof the whole population.



60

9.

Sampling error - The difference betweenthe results obtained from a survey sampleand those that would have been obtainedif the entire population was surveyed. Thesize of sampling error varies both with thesize of the sample and with thepercentages giving a particular response.

Screening - The practice of distinguishingbetween individuals who should beenrolled in a program/intervention andthose who should not be enrolled; a toolfor identifying individuals at risk; toexamine carefully to determine suitability.

Self-selection - The act wherebyindividuals determine their participationin some activity or event, because ofunderlying values, characteristics orcircumstances.

Simple cross-section survey - A surveythat collects data using random sampling;a survey that gives all individuals orhousehold in the study area an equalchance of being chosen for the survey.

Simple random sample - The results of amethod of sampling that gives everyonean equal chance of being selected; thesimplest form of probability sampling; asample in which an individual’s selectionis independent of the selection of anyother individual.

Skewed distribution - A distribution inwhich one side is unequal to the otherside; a distribution in which the two sidesdo not mirror each other across the centerline of the mean.

Skinfold calipers - A tool used to assessskinfold thickness by measuring thethickness of the skin pinched between itsprongs.

Skinfold thickness - A measurement thatprovides an estimate of subcutaneous fatdeposits, which in turn providesinformation on total body fat.

Specificity - Characteristic of aclassification system that correctlyidentifies children who are not at risk; theprobability that a healthy individual willbe classified as healthy; a system with fewfalse positives.

Spring scale - A scale that measuresweight by the amount a spring is pulled bythe object being weighed; a hanging scale.

Standard deviation - A statisticalmeasure of dispersion away from themean; the square root of the variance.

Stratified sample - A method of samplingthat ensures proportional representationfrom all sub-groups or strata.


9.

61

Stratified survey - A survey that choosesparticipants randomly after they havebeen divided into the applicable strata orsub-groups.

Student’s t-test - A statistical test todetermine if there is a significantdifference in means of a continuousvariable between two groups.

Stunting - A slowing of skeletal growththat results in reduced stature or length; acondition that usually results fromextended periods of inadequate foodintake, especially during the years ofgreatest growth for children.

Subcutaneous fat - Fat located justunderneath the skin; fat that is used as ameasure of total body fat stores in skinfoldthickness measurements.

Subscapular area - The site just belowthe shoulder blade; situated below or onthe underside of the scapula.

Summary statistics - Statistics that areused to describe the center and spread ofthe distribution of a variable; statisticsthat usually make up such a summaryinclude: the mean, standard deviation,median, variance, mode, total, standarderror, and upper and lower quartiles.

Survey - A method of gatheringinformation about a large number of peopleby talking to a few of them; a way to collectinformation on people’s needs, behavior,attitudes, environment and opinions, aswell as on such personal characteristics asage, income and occupation.

Systematic sample - A modification of asimple random sample, that consists ofpicking individuals at regular intervalsfrom a random list.

Tolerated sampling error - The amountof difference permissible between theestimate (results from a survey sample)and the actual value in the population.

Trend assessment - The process oftracking nutritional progress over time;examples of nutritional trend assessmentinclude growth monitoring andnutritional surveillance.

Trend studies - A type of longitudinalsurvey that uses different individuals forstudy over time; a study which usesdifferent households in each survey, butin which each sample represents thesame general population at differenttimes.

Triceps - The muscle at the back of theupper arm; a large three-headed musclerunning along the back of the upper armand functioning to extend the forearm.



62

9.

Two-tailed Test - A statistical test to detecta difference in means between twopopulations regardless of the direction ofthe difference.

Underweight - A condition measured byweight-for-age; a condition that can alsoact as a composite measure of stunting andwasting.

Variable - A quantity that may vary fromobject to object; a characteristic of a unit.

Wasting - A condition measured byweight-for-height; a condition that resultsfrom the loss of both body tissue and fat, ina body; a condition that usually reflectsseverely inadequate food intake happeningat present.

Waterlow classification system - Anutritional classification system that usespercentage of the median of height-for-ageand weight-for-height in combination toidentify children who are wasted, stuntedor both.

Weighing trousers - A pair of little pantsthat a child can step into and be suspendedby for weighing from a hanging scale.

Weight-for-age - An index of acutemalnutrition; a valuable index for use withvery young children or when lengthmeasurements are difficult to doaccurately.

Weight-for-height - An index of currentnutritional status also referred to aswasting.

Weighting - A data analysis process thatinvolves adjusting key variables used forsample selection to their actualproportions in the population.

Z-score - A statistical measure of thedistance, in units of standard deviations, ofa value from the mean; the standardizedvalue for an item based on the mean andstandard deviation of a data set; astandardized value computed bysubtracting the mean from the data value xand then dividing the results by thestandard deviation.


63

Acronyms

BHR Bureau of Humanitarian ResponseBMI Body Mass IndexCDC Centers for Disease Control and PreventionCm CentimetersCS Cooperating SponsorDHS Demographic and Health SurveyDHS-III Demographic & Health Survey (third phase DHS surveys conducted in country)DOB Date of birthDOM Date of measurementEBF Exclusive breastfeedingFFP Food for PeaceGMP Growth monitoring and promotionHAZ Height for age Z-scoreHAM Height for age % medianHAP Height for age percentileHt HeightID IdentificationKgs KilogramsKPC Knowledge Practice and CoverageMCH Maternal and child healthMUAC Mid-upper arm circumferenceNGO Non-governmental organizationPVO Private voluntary organizationUNICEF United Nations Children’s FundUSAID United States Agency for International DevelopmentSPSS Statistical Package for Social Sciences (software)USAID United Stated Agency for International DevelopmentWAZ Weight for age Z-scoreWAM Weight for age % medianWAP Weight for age PercentileWHZ Weight for height Z-scoreWHM Weight for height for age % medianWHP Weight for height percentileWHO World Health OrganizationWt Weight



64

9.

Using Distributions by Standard DeviationThe reference population data areavailable with the mean measurementvalues and the measurement values for upto +/- 3SD displayed, for the threenutritional status indices. These valuesare given for each month of age up to 119months, for both boys and girls. The datamay be viewed in the following threeformats: (1) table; (2) graph; and (3)spreadsheet. An example of how a table ofweight-for-age values for boys can be usedis as follows.

ProcedureA 19-month-old boy who weighs 9.8 kg iscompared to the reference standards. Use adistribution of the reference standards thatalready has values for the standarddeviations calculated. First, the appropriatedistribution should be consulted.

Example: Part of the table below is theweight-for-age by standard deviation forboys. Age in months is listed in the far leftcolumn. The mean or expected values foran “average” healthy boy of each age islocated in the middle column. Themeasurement values range from -3 to +3SDs, with the standard deviation for thelower and upper halves of the distributionalso shown. The correct line of the tableshould be found (lines vary by months ofage or centimeters of length). The child’smeasurement should then be pinpointedalong this line.

Example: In this case, we need to locatethe line for the age 19 months and thenfind where 9.8 kg falls. The table showsthat such a child falls between -2 and -1SD.

Therefore, we would say that a 19 monthold boy who weighed 9.8 kg is between -2and -1 SD from the mean. To obtain amore accurate statistic, a Z-score wouldhave to be calculated.

Calculating Z-scoresWhen the mean and standard deviationfor a set of data are available, as they arewith the reference standards, a Z-scorecan be calculated. In this case, a Z-scorecalculated for an individual tells exactlyhow many standard deviation units hismeasurements are away from the mean ofthe reference distribution. A positive Z-score means that an individual’smeasurements are higher than thereference mean and a negative Z-scoremeans that the measurements are lowerthan the reference mean. The advantageof calculating a Z-score is that it providesmore precision than just locating aposition on a table, as we did above. Onlythe mean and standard deviation areneeded.

ProcedureAssume we have the same 19 month oldboy from the example above, who weighs9.8 kilograms. If we look at the reference

1. CalculatingZ-scoresA

PP

EN

DI

X


9.

65

Weight (kg) for age distribution by standard deviation for boys

Age Group(months)

12

13

14

15

16

17

18

19

20

21

22

23

24

LowerS.D.

1.0

1.0

1.1

1.1

1.1

1.1

1.2

1.2

1.2

1.3

1.3

1.3

1.1

-3S.D.

7.1

7.3

7.5

7.6

7.7

7.8

7.9

8.0

8.1

8.3

8.4

8.5

9.0

-2S.D.

8.1

8.3

8.5

8.7

8.8

9.0

9.1

9.2

9.4

9.5

9.7

9.8

10.1

-1S.D.

9.1

9.4

9.6

9.8

10.0

10.1

10.3

10.5

10.6

10.8

10.9

11.1

11.2

Mean

10.2

10.4

10.7

10.9

11.1

11.3

11.5

11.7

11.8

12.0

12.2

12.4

12.3

+1S.D.

11.3

11.5

11.8

12.0

12.3

12.5

12.7

12.9

13.1

13.3

13.5

13.7

14.0

+2S.D.

12.4

12.7

13.0

13.2

13.5

13.7

13.9

14.1

14.4

14.6

14.8

15.0

15.7

+3S.D.

13.5

13.8

14.1

14.4

14.7

14.9

15.2

15.4

15.6

15.8

16.0

16.2

17.4

UpperS.D.

1.1

1.1

1.2

1.2

1.2

1.2

1.2

1.2

1.3

1.3

1.3

1.3

1.7

standards for weight-for-age we see thatthe “average” healthy boy of 19 monthsshould weigh 11.7 kilograms. Since thischild is obviously under the mean of 11.7kilograms, we need to check the lowerstandard deviation value (remember: withweight-for-age the lower and upperstandard deviations might differ!). It is 1.2kilograms. With these two pieces ofinformation, we can calculate the Z-scoreof a child’s weight-for-age, using thefollowing procedures:

Subtract the mean weight from the actualweight of the child. The results in this casewill be negative.

Example: 9.8 kg - 11.7 kg = -1.9

Divide the result by the standard deviationfor the child’s age and gender.

Example: -1.9 / 1.2 sd = -1.58 SD units

The resulting number is the Z-score forthat child.

Example: The Z-score for a 19 month oldboy who weighs 9.8 kg is -1.58 standarddeviation units.

This procedure can be repeated with theappropriate graphs or tables to calculate

the Z-scores for length-for-age andweight-for-height.

Percent of the Median

ProcedureFor this example, we will assume we havejust measured the length of a girl who is24 months old. This girl is 64.9 cm long.We want to use percentages of the medianto compare her to the reference standards.Using the percentile distribution for theappropriate index (weight-for-age, length-for-age or weight-for-length) and sex, findthe measurement that corresponds to the50th percentile (remember the 50thpercentile is the same as the median).

Example: From a table of referencevalues we learn that the 50th percentilelength measurement for a 24 month oldgirl is 86.5 cm.

Divide the measurement of the individual childby the appropriate median measurement.

Example: Our girl is 64.9 cm. The median value for girls is 86.5 cm.

64.9 cm / 86.5 cm = .75 Multiply this fraction by 100, to convert itto a percentage.Example: 0.75 x 100 = 75%



66

9.

A 24 month old girl who is 64.9 cm long is75 percent of the median. The procedurewould be the same for a child who has ameasurement that is larger than themedian. That child would, however, beover 100 percent of the median.

The main disadvantage of this system isthe lack of exact correspondence with afixed point of the distribution across age orheight status. For example, depending onthe child’s age, 80% of the median weight-for-age might be above or below -2 Z-scores. In terms of health, beingbelow or above -2 Z-scores this wouldresult in a different classification of risk.In addition, typical cut-offs for percent ofmedian are different for differentanthropometric indices.

To approximate a cut-off of -2 Z-scores,the usual cut-off for low height-for-age is90%, and for low weight-for-height andlow weight-for-age 80% of the median.

The table below is the reference forlength-for-age for boys aged 0-12 months.The value given in the “Mean” column isthe average length in centimeters that wewould expect for a healthy boy at each age.Age in months is given in the far left-handcolumn. The column next to it shows thenumber of centimeters needed to equal 1

standard deviation unit (you should noticethat the number of centimeters in onestandard deviation unit generallyincreases as age goes up).

In the example of a 3 month old boy,we can see under the mean column in thetable, that 61.1 cm is the expectation for ahealthy boy. If we had measured a 3month old boy who was 2.6 cm under thisexpectation (or 58.5 cm), we would findthat this measurement falls exactly 1standard deviation unit under the mean.Therefore, we could state that thisindividual child is -1 standard deviationfrom the mean or average (expectation)for length-for-age. A 12 month old boywho measured 2.7 cm under theexpectation for his age (or 73.4 cm), wouldalso be 1 standard deviation unit underthe mean length-for-age, for his age.Standard deviation provides an easy wayto tell what measurements are of equalconcern when we are measuring boys andgirls of different ages.

Length (cm) for age (months) distribution by standard deviation for boys

Age Group(months)

0

1

2

3

4

5

6

7

8

9

10

11

12

LowerS.D.

2.3

2.5

2.6

2.6

2.7

2.7

2.7

2.7

2.7

2.6

2.6

2.7

2.7

-3S.D.

43.6

47.2

50.4

53.2

55.6

57.8

59.8

61.5

63.0

64.4

65.7

66.9

68.0

-2S.D.

45.9

49.7

52.9

55.8

58.3

60.5

62.4

64.1

65.7

67.0

68.3

69.6

70.7

-1S.D.

48.2

52.1

55.5

58.5

61.0

63.2

65.1

66.8

68.3

69.7

71.0

72.2

73.4

Mean

50.5

54.6

58.1

61.1

63.7

65.9

67.8

69.5

71.0

72.3

73.6

74.9

76.1

+1S.D.

52.8

57.0

60.7

63.7

66.4

68.6

70.5

72.2

73.6

75.0

76.3

77.5

78.7

+2S.D.

55.1

59.5

63.2

66.4

69.1

71.3

73.2

74.8

76.3

77.6

78.9

80.2

81.5

+3S.D.

57.3

61.9

65.8

69.0

71.7

74.0

75.9

77.5

78.9

80.3

81.6

82.9

84.2

UpperS.D.

2.3

2.5

2.6

2.6

2.7

2.7

2.7

2.7

2.7

2.6

2.6

2.7

2.7


9.

67

2.Uses ofAnthropometric Data

AP

PE

ND

IX

Anthropometric indicators can beclassified according to the objectives oftheir use, which include the following (theorder of listing is dictated by variousmethodological considerations discussedlater):

Identification of individuals orpopulations at risk. In general, thisrequires data based upon indicators ofimpaired performance, health or survival.Depending on the specific objective, theanthropometric indicators should:

• reflect past or present risk, or

• predict future risk

An indicator may reflect both present andfuture risk; for instance, an indicator ofpresent malnutrition may also be apredictor of an increased risk of mortalityin the future. However, a reflectiveindicator of past problems may have novalue as a predictor of future risk; forexample, stunting of growth in earlychildhood as a result of malnutrition maypersist throughout life, but with ageprobably becomes less reliably predictiveof future risk.

Indicators of this type might be used inthe risk approach to identification ofhealth problems and potentialinterventions, although, the risk approachmay have little value in predicting orevaluating the benefit derived frominterventions.

Selection of individuals or populationsfor an intervention. In this application,indicators should predict the benefit to bederived from the intervention.

The distinction between indicators ofrisk and indicators of benefit is not widelyappreciated, yet it is paramount fordeveloping and targeting interventions.Some indicators of present or future riskmay also predict benefit, but this is notnecessarily the case. Low maternalheight, for example, predicts low birthweight, but, in contrast to low maternalweight in the same population, does notpredict any benefit of providing animproved diet to pregnant women. By thesame token, predictors of benefit may notbe good predictors of risk.

Anthropometry provides importantindicators of overall socioeconomicdevelopment among the poorest membersof a population. Data on stunting inchildren and adults reflects socioeconomicconditions that are not conducive to goodhealth and nutrition. Thus stunting inyoung children may be used effectively totarget development programs.

Evaluation of the effects of changingnutritional, health, or socioeconomicinfluences, including interventions. Forthis purpose indicators should reflectresponses to past and presentinterventions. This is the case with Title IIprogram evaluations.



68

9.

Change in weight-for-height (wasting) is agood example of an indicator of short-term response in a wasted child beingtreated for malnutrition, whereas adecrease in the prevalence of stunting atthe population level is a long-termindicator that social development isbenefiting the poor as well as thecomparatively affluent. On the otherhand, a decrease in the prevalence of lowbirth weight might be used to indicatesuccess in such activities as controllingmalaria during pregnancy.

In describing an indicator of response,the possible lag between the start of anintervention and the time when aresponse becomes apparent is animportant consideration. At the individuallevel, a wasted infant will respond toimproved nutrition first by putting onweight and then by “catching up” in lineargrowth. At the population level, however,decades may elapse before improvementscan be seen in adult height.

Excluding individuals from high-risktreatments, from employment or fromcertain benefits. Decisions regarding anindividual’s inclusion in, or exclusionfrom, a high-risk treatment protocol,consideration for employment in aparticular setting (e.g. an occupationrequiring appreciable physical strength),or admission to certain benefits (e.g. lowlife-insurance rates) depend on indicatorsthat predict a lack of risk.

Anthropometric indicators of lack ofrisk were once presumed to be the sameas those that predict risk, but recent workhas revealed that this is not invariably thecase. Studies have found that indicators ofpoor growth were less effective inpredicting adequate growth than otherindicators.

Achieving normative standards.Assessing achievement of normativestandards requires indicators that reflect“normality”.

Some activities appear to have noobjectives beyond encouraging indi-viduals to attain some norm. For instance,some have argued that moderate obesityamong the elderly is not associated with

poor health or increased risk of mortality,and if this were true, advocacy for theneed for weight control in this age groupwould be based solely on normativedistributions.

Research purposes that do not involvedecisions affecting nutrition, health, orwell-being. The indicator requirementsfor these objectives, whether they concernindividuals or whole populations, aregenerally beyond the scope of this guide.

There may be differences in theinterpretation of anthropometricindicators when applied to individualsor to populations. For example, while areflective indicator, such as the presenceof marasmus, signifies malnutrition in agiven child today, a sudden increase ofmarasmus in a population may bepredictive of future famine.

The appropriateness of indicators thusdepends on the specific objectives of theiruse, and research is only just beginning toaddress this specificity and itsimplications. Little is known, for example,about how the use of different cut-offs foranthropometric indicators fulfills differentobjectives.


9.

69

3.Selectinga Sample

AP

PE

ND

IX

Sampling is the process that is used toselect a representative group of individualswhose characteristics can be described andused to represent the whole population.The following section outlines a few of theproblems that can result from incorrectsampling.

Sampling ErrorA lot of thought should go into how toselect the sample, and the followingexamples illustrate this.

Example: You are interested in doing asurvey that will gather information on theprevalence of underweight childrenbetween 6 and 35 months and on thecharacteristics of their households. Eachvillage that you will be conducting surveysin has a clinic where mothers and theirchildren can be found on certain days ofthe week. The clinics already have scales,which would make them convenient. Youdecide to go to the clinics and measure allthe children found there. Is this a goodsampling plan for conducting your survey?

The answer is no. Mothers who go tothe clinic are considered self-selecting.The mothers who decide to go to the clinicmight have reasons for going that wouldmake them different from mothers who donot go to the clinic. They might be moreconcerned with their child’s health. Orthey might be mothers who have sickerchildren. Or they might be mothers whodo not have to work everyday and

therefore have the time to take theirchildren to the clinic.

Any characteristic that makes anindividual do something may also makethem different from the population as awhole and have an effect on theirchildren’s nutritional status. When youhave a sample that is made up of peoplewho are not representative of thepopulation for a certain reason, you havewhat is called sampling bias. This type ofsampling error can happen when youselect a sample from a group who all go toa specific place, but it can also happen ifthe sample is from one area of a village ora city or among people who are neighbors.These individuals may share somecharacteristic that makes them moresimilar to each other than to the largerpopulation.

Here are some tips for avoiding this type ofsampling error:

• Do not choose samples exclusivelyfrom particular groups, such as childrencoming to clinics.

• Do not ask mothers to bring theirchildren to a central point in thecommunity, because some of them willnot come; you will not be able to find outhow many failed to appear and howdifferent they may be from those whocame.

• Do not use samples chosen at will bythe interviewer, field supervisor or fielddirector.



70

9.

• Do not restrict your sample to familiesliving in easily accessible households,such as those close to a main road or neara village center; families living in lessaccessible areas may be poorer and lesshealthy.

• Do not omit households where no oneis at home the first time you call.

Sampling error can also happen if thesample size is not large enough.

Sample Size

Factors influencing sample size decisions

The sample size required for a givensurvey is determined by its measurementobjectives. For surveys designed to eithermeasure changes in indicators over timeor differences in indicators betweenproject and control areas, the requiredsample size for a given indicator for eachsurvey round and/or comparison groupdepends upon five factors:

• how numerous the measurement unitsfor the indicator are in the targetpopulation;

• the initial or “baseline” level of theindicator;

• the magnitude of change orcomparison group differences on theindicator it is desired to be able to reliablymeasure;

• the degree of confidence with which itis desired to be certain that an observedchange or comparison-group difference ofthe magnitude specified above would nothave occurred by chance (that is, the levelof statistical significance); and

• the degree of confidence with which itis desired to be certain of measuring anactual change or difference of themagnitude specified above will bedetected (i.e., statistical power).

Note that the first two of these parametersare population characteristics, while thelast three are chosen by the evaluator/survey designer.

An example using changes innutritional status:

To illustrate how these parameters enterinto the determination of sample sizerequirements, consider an evaluationwhere changes in indicators for theproject area are being measured over time(i.e., a one-group pretest-post-test orreflexive design). For such an evaluation,the objectives for sample sizedetermination purposes might be stated interms of a key indicator as follows: to beable to measure a decrease of 20percentage points in the proportion ofchildren 6-59 months of age who arestunted with 95% confidence and 80%power. Thus, if the estimated proportionof children who were stunted at the timeof the baseline survey was 40%, theobjective would be to measure a change inthe prevalence of stunted children from40% to 20% and be (1) 95% confident thata decline of this magnitude would nothave occurred by chance and (2) 80%confident of detecting such a decline if oneactually occurred. The sample sizecalculations would answer the questions(1) how many children ages 6-59 monthswould be required to accomplish theabove objectives and (2) how manyhouseholds would have to be chosen inorder to find this number of children.

For evaluation designs involving com-parisons between project and controlareas, the objectives are framed in termsof the magnitude of differences betweenthe two groups it is desired to be able toreliably detect. For example, in a reflexiveevaluation design, sample size require-ments might be set to detect a differencebetween project and control areas of 20percentage points on a specified indicator.Similarly, when a pre- and post-testdesign with treatment and control areas isto be used, the sample size would be set toensure that a difference in the degree ofchange on a key indicator between projectand control areas of a specified magnitude(e.g., 20 percentage points) could bereliably detected.


9.

71

Figure A3.1. Illustrative informational needs for determining sample size, genericTitle II "health" indicators

A. Information on population composition: 1. Mean number of persons per household 2. Proportion of total population that are: a. Children under 0-59 months of age. b. Children under 24 months of age. c. Infants under 6 months of age. d. Infants between the ages of 6 and 10 months.

B. Information about "expected" levels or rates in the target population: 1. Proportion of children aged 6-59 months who are stunted. 2. Proportion of children aged 6-59 months who are underweight.

Steps involved in determining survey sample size requirements for a given survey:

• Calculating the number of sample elements required in order to satisfy the measurement requirements for a given indicator, and how many households would have to be contacted in order to find the number of elements needed in the first step.

a. For indicators expressed as proportions:

The following formula may be used to calculate the required sample size for indicators expressed as a percentage or proportion. Note that the sample sizes obtained are for each survey round or each comparison group:

alpha = ∞ beta = ßn = D [(Z∞ + Zß)2 * (P1 (1 - P1) + P2 (1 - P2)) /(P2 - P1)2]

or n = D [(Z alpha + Z beta) squared times (P one times (1 minus P one) plus P two times (1 minus P two)) divided by (P two minus P one) squared]

Where:

n = required minimum sample size per survey round or comparison group;

D = design effect;

P1 = P one, the estimated level of an indicator measured as a proportion at the time of the first survey or for the control area;

P2 = P two, the expected level of the indicator either at some future date or for the project area such that the quantity (P2 - P1) is the size of the magnitude of change it is desired to be able to detect;

Z∞ = Z alpha, is the Z-score corresponding to the degree of confidence with which it is desired to be able to conclude that an observed change of size (P2 - P1) would not have occurred by chance; and

Zß = Z beta, is the Z-score corresponding to the degree of confidence with which it is desired to be certain of detecting a change of size (P2 - P1) if one actually occurred.

* refers to a multiplication.Standard values of Z alpha (Z∞ ) and Z beta (Zß ) are provided in Figure A5.2, and the use of the above formula is illustrated in Figure A5.3. The different parameters in the formula are discussed below.



72

9.

Figure A3.2. Values of Z alpha (Z∞) and Z beta (Zß)

alpha ∞.90

.95

.975

.99

One-tailed test

1.282

1.645

1.960

2.326

Two-tailed test

1.645

1.960

2.240

2.576

beta ß

.80

.90

.95

.975

.999

Z beta Zß

0.840

1.282

1.645

1.960

2.320

Figure A3.4. Sample sizes required for selected combinations of P one (P1) and changesor comparison-group differences to be detected (for alpha (∞) = .95 and beta (ß) = .80)

Change/difference to be detected (P2 - P1) (P two minus P one)

P one

Figure A3.3. Illustrative sample size calculations for indicators expressed as proportions

Example 1

Suppose that it were desired to measure a decrease in the prevalence of underweight (weight-for-age) of 10 percentage points. At the time of the first survey, it is thought that about 40 percent of children between 12 and 36 months were underweight. Thus, P1 = .40 and P2 = .30. Using ‘standard’ parameters of 95 percent level of significance and 80 percent power, values from Figure A5.2 of alpha (∞) = 1.645 (for a one-tailed test - see below for further discussion) and beta (ß) = 0.840 are chosen. Inserting these values into the above formula, we obtain:

n = 2 [(1.645 + 0.840) 2 * ((.3)(.7) + (.6)(.4))] / (.3 - .4) 2

= 2 [(6.175 * 0.45)] / .01

= 2 * [2.77875] / .01 = 2 (277.875) = 555.75

or 556 households per survey round.

Figure A3.4 provides a "lookup" table based upon the above formula to permit sample sizes to be chosen without having to perform calculations. The table provides sample sizes needed to measure changes/differences in a given indicator of specified magnitudes P two minus P one (P2 - P1) for different initial levels of the indicator (P1). The table is for values of alpha (∞) = 0.95 and beta (ß) = 0.80.

Note: sample sizes shown assume a design effect of 2.0 and one-tailed tests. In a study of population-based cluster surveys to determine the design effects Katz (AJCN, 1995 Jan; 61(1):155-60) found the design effect range from 0.44 to 2.59. The use of D=2.0, therefore is conservative. For values of P one (P1) greater than .50, use the value in the table that differs from .50 by the same amount. For example, for P one (P1 ) = .60, use the value for P one (P1 ) = .40; for P one (P1 ) = .70, use the value for P one (P1 ) = .30.

P1

.10

.15

.20

.25

.30

.35

.40

.45

.50

.05

1,075

1,420

1,176

1,964

2,161

2,310

2,408

2,458

2,458

.10

309

389

457

513

556

587

606

611

606

.15

152

185

213

235

251

262

268

268

262

.20

93

110

124

134

142

147

148

147

142

.25

63

73

81

57

90

92

92

90

87

.30

45

52

56

60

62

62

62

60

56


9.

73

For indicators expressed as means or totals

For indicators that are means or totals, the following formula may be used to calculate sample size requirements for each survey round or comparison group:

n = D [(Z∞ + Zß)2 * (sd12 + sd22) / (X2 - X1)2]

(n equals D time [(Z alpha plus Z beta) squared times (sd one squared + sd two squared) divided (X two minus X minus) squared]

Where:

n = required minimum sample size per survey round or comparison group;

D = design effect; Z∞ = Z alpha the Z-score corresponding to the degree of confidence with which it is desired to be able to conclude that an observed change of size (X2 - X1) would not have occurred by chance;

Zß = Z beta the Z-score corresponding to the degree of confidence with which it is desired to be certain of detecting a change of size (X2 - X1) if one actually occurred; sd1 and sd2 = "expected" standard deviations for the indicators for the respective survey rounds or comparison groups being compared;

X1 = X one is the estimated level of an indicator at the time of the first survey or for the control area; and

X2 = X two is the expected level of the indicator either at some future date or for the project area such that the quantity (X2 - X1) is the size of the magnitude of change or comparison-group differences it is desired to be able to detect.

The primary difficulty in using the above formula is that it requires information on the standard deviation of the indicator being used in the sample size computations. The preferred solution to this problem would be to use values from a prior survey that had been undertaken in the setting in which a program under evaluation is being carried out. If such data are not available, data from another part of the country or a neighboring country with similar characteristics may be used. Such data are often presented in survey reports.



74

9.

4. MeasuringAdultsA

PP

EN

DI

X

The use of indicators of adult nutritionalstatus for evaluating and monitoringUSAID Title II development programs islimited. There is some experience withmicronutrient status including anemiaand Vitamin A and more with assessingbody mass index for women. This sectiondoes not deal with obesity, micronutrientmalnutrition or pregnancy. The inform-ation in this section refers to anthro-pometric assessment and is derived fromvarious sources including FAO, WHO andthe recent publication from the ACC/SCNavailable from their website:acc.unsystem.org/scn/.

Adults are defined by WHO as those inthe age range of 25-60 years althoughcategories often extend from 20 years to 65years of age. Adult anthropometrics havenot been standardized in terms ofreference data or choice of indicators forrisk and response assessment as they havebeen for children. As noted in varioussources, there is no recommendedindicator or assessment approach for adultnutritional status. The assessment ofadults older than 60 years presents anumber of specific challenges not coveredhere. The reader is referred to the work ofHelp Age to deal with these assessments(Ismail, S. and Manandhar, M. 1999).

As with children, adult anthropometricassessment is used to reflect under-nutrition. Anthropometry is also used toreflect over nutrition but this is not thefocus of this guide. Undernutrition in

adults is characterized by patterns of acuteand chronic deficiency of energy, proteinand micronutrients including vitaminsand minerals. Often a person is affectedby both acute and chronic deficiency in allor some of the key nutrients. Themanifestation of the deficiency and themeasurement is therefore, complicated todetermine and the functional significanceunclear. Undernutrition is characterizedby a lack of food and while specificnutrient deficiencies occur, such aspellagra due to a lack of niacin, theprimary cause is more general. We arelearning more about specific nutrientrequirements for diseases such as HIV/AIDS but the ability of anthropometrics toidentify these conditions is limited.

Adult anthropometric assessment is usedfor several purposes including:

• screening or targeting individuals forsome sort of intervention or action such assupplementary feeding during faminerelief,

• surveillance or monitoring of changesin prevalence and coverage in groups orpopulations to trigger a responseincluding graduating from an intervention,

• evaluating the impact of activities orinterventions.

Anthropometry is used to describe thenutritional situation in a population andthis can be useful for problem analysisand for evaluation. Because the deter-


9.

75

minants of nutrition are so many, it isimportant to examine other factors thanjust anthropometry such as the foodsecurity situation, levels of illness, caregiving practices and so on.

For assessing women’s nutrition status,usually, a combination of indicators isneeded. Unlike children, reference datahave not been standardized for women. Forcross-sectional comparisons, referencedata can be gathered from within the samepopulation that the intervention groupbelongs to among healthy women, womenwith positive pregnancy outcomes.Construction of an adult nutrition referencehas been done for some populations.

Each anthropometric indicator listed foradults and especially for women has itsown advantages and limitations. Todetermine the best indicator, one shouldconsider the objectives of the nutritionprogram and its associated reportingrequirements. Several possible uses ofindicators are:

4.1 HeightAdults are not growing in stature. Heightin adults is determined by a person’sgenetic potential and the health andnutrition experiences dating back to thefetus. Most growth in height is completedby age 17 with some incremental growthfor another 10 years. Height, therefore,may be useful for a reflection of pastevents and be used in some screeningsituations, but will not be able to reflectrecent or current nutritional shocks orchange. Women’s height is a usefulpredictor of pregnancy outcomes such aslow birth weight and possible deliverycomplications; thus it is an indicator ofrisk. Because height will change very littleamong adult women, it is not useful forevaluating interventions (outcome).Height can not be used for monitoring andevaluation of programs.

Adult Measuring Device (Microtoise) -(UNICEF No. 0114400 Height measuringinstrument (0-2 m)) This lightweightportable tape is wall mounted and fitseasily into the package needed for field

measurements. Made of plastic, theMicrotoise measures up to 2 meters and isavailable for approximately $US20. Referto the local UNICEF office forprocurement options.

Adult Measuring Device - Aninexpensive height measuring deviceuseful for children over 24 months andadults is the Harpenden pocketstadiometer (range 0-2m) available fromCMS weighing equipment. The cost isapproximately $US100. CMS WeighingEquipment, Ltd., 18 Camden High School,London NW1 OJH, U.K.; Telephone: 01-387-2060 (international) +44 171 387 2060.

4.2 WeightWeight does change reflecting recentevents. For monitoring change in anindividual, weight change is helpful asheight would not be expected to change.For controlled studies where the sameparticipants are being followed and heightwould not be changing, weight can beused for screening and evaluationpurposes. More than one measurement isnecessary for tracking changes.

For women who are pregnant, weightgains of 1.5 kg/month during the last twotrimesters are consistent with positivepregnancy outcomes in developingcountries. Short maternal stature, lowpregnant BMI and poor weight gainduring pregnancy are all indicators of riskfor low birth weight. As stated above, notall of these are indicators of outcome.

For comparing within or acrossgroups, an interpretation of weight changeneeds to be done controlling for thevariation in height. The most commonindicator used to control for height and toreflect body mass is referred to as the bodymass index (BMI). There are limitationsin the use of this indicator.

Scales are used to weigh adults and can beobtained from various sources. Ensurethe scale is sturdy, reliable and accurate.



76

9.

UNICEF Electronic Scale - (Item No.0141015 Scale mother/child, electronic)The scale is manufactured by SECA and isa floor scale for weighing children as wellas adults (capacity 150 kg). Weighingcapacity from 1 kg to 150 kg in 100 gdivisions, accuracy +/- 100 g. Weight ofadult on scale can be stored (tared) inmemory, allowing weight of baby or smallchild held by adult to show on scaleindicator. Solar cell on-switch (lightsensitivity 15 lux). Powered by long-lifelithium battery, good for one millionweighing cycles. Portable, weight 4 kg.Instructions in English, French andSpanish.

The major advantage is that it has amicro-computer chip so that it can adjustitself to zero and weighs people quicklyand accurately. The child may be weigheddirectly, if possible. If a child is frightened,the mother can first be weighed alone andthen weighed while holding the child inher arms, and the scale will automaticallycompute the child’s weight by subtraction.Recent experience in surveys suggests thatthe scale is appropriate for CooperatingSponsor use although some difficulty hasbeen experienced with heat adverselyaffecting the scale. The price of this scaleis US$90 and it can be ordered fromUNICEF's Supply Division in Copenhagenthrough any UNICEF field office.www.supply.unicef.dk/catalogue/index.htm

BMI is based on a weight-to-height ratiothat is considered a good index of body fatand protein stores. Body stores are ofinterest because they reflect the storesneeded to cope with physiological stressdue to reduced intake and increaseddemands due to increased activity,pregnancy and diseases. Adults who havea healthy nutritional status would beexpected to have body stores or BMIwithin a certain range. BMI - also knownas "Quetelet’s index” - is summarizedbelow:

Body mass index (BMI) = weight/(height)2

The formula for BMI is the weight (inkilograms) divided by the height (inmeters) squared. A woman who weighs

55.5 kgs and a height of 162.5 cm wouldhave a BMI of ((55.5/(1.625 x 1.625)) =20.9.

It is best used for individuals between theages of 20 and 65 years.

While no standard classification systemexists, the following was recommended byCollins et al. (2000) for chronicallyundernourished populations.

Table A4.1. Body Mass Index Categories of Chronic Undernutrition

Normal >= 18.5

Grade I 17.0 – 18.4

Grade II 16.0 – 16.9

Grade III <= 15.9

While these categories are suggested, thereis difficulty of using these categories tocompare across populations due to 1) a lackof understanding of the functionalsignificance of these categories and 2) theinfluence of body shape to interpreting BMI.

Body shape, especially the size of thetrunk in relation to the leg lengthinfluences both the BMI and theinterpretation of the result. A long wasted– short legged individual will have ahigher BMI for the same overall height assomeone with especially long legs. Bodyshape can be reflected in the ratio ofsitting height (reflecting the trunk length)to the standing height (reflecting the leglength). Referred to as the Cormic Index(sitting height to standing height,expressed as a percentage – SH/S), thisindex can correct for differences in BMI inethnically diverse populations. The SH/Sindex should be expressed as apercentage.

The SH/S percentage can be measuredusing the standing height (H) (see above)and the sitting height (SH) measured bythe person sitting upright in a chair eitherin a measuring board used for children(refer to Section 4) or the adult measuringdevices discussed above.


9.

77

The correction from Collins et al. (2000)for standardizing BMI using the CormicIndex (SH/S percentage) applies thefollowing formula:

Male BMI = 0.78 (SH/S) – 18.43

Females BMI = 1.19 (SH/S) – 40.34

The reader is referred to Collins et al.(2000) for the use of the Cormic Index forcomparing across population groupswhere an average for the Cormic Indexcan be used. For individuals, their ownCormic Index should be used to correctthe BMI.

BMI changes during pregnancy. It will benecessary to separate pregnant womenfrom non-pregnant women when com-paring BMI.

In summary:

• For comparisons within populationsover relative short times for evaluationpurposes, BMI should not require acorrection using the Cormic Index.

• Assessing or screening individuals fortargeting using BMI should correct theindividuals BMI using the Cormic Indexfor that individual.

The authors of the ACC/SCN review(Collins et al., 2000) conclude that insituations of emergency screening, themeasurement of height and sitting heightand the use of the Cormic Index correctionis time consuming and mostly unrealistic.In situations of famine relief, they cautionthat without the standardization with theCormic Index, the use of BMI alone forscreening is inappropriate. BMImeasured during emergency situationsrequires good equipment, well-trainedpersonnel and an ability to convert themeasures into the BMI. Challenges inmeasuring people who are very sick,elderly and disabled may make the use ofBMI for screening even more difficult.

In addition to a Cormic Index correction,the age distribution is important. Aspeople grow older, the distribution of fat

and fat free mass (water, bone andmuscle) changes. For screening purposes,BMI may have to be adjusted for age.There is no guidance for this at themoment. For evaluation purposes, as longas the distribution of age in the baselineand follow up remain the same, the biaseswill be consistent and age should notconfound the analysis of change. Whereage is a potential confounding variable,data may have to be presented for agegroup which has assessment implications(not all respondents know their ages) andsample size estimations (stratification ofthe sample increases size requirements).

4.3 Mid-upper arm circumference(MUAC)MUAC is the circumference of the leftupper arm measured in centimeters. Thepoint is between the tip of the shoulderand the elbow. The use of MUAC and theequipment for the measurement isdetailed in section 5 of this Guide. Armcircumference is measured with specialcircumference measuring tapes orcircumference insertion tapes.

Because the equipment is lightweightand training to do MUAC isstraightforward, MUAC is utilized forscreening in emergency situations whennutritional status information is neededimmediately for large groups of people,especially children. The indicator is usefulfor assessing acute adult undernutritionand for assessing the prevalence ofundernutrition at the population level.

While arm circumference measuresboth muscle and fat, some populationswould be expected to have very littlesubcutaneous fat on their arms. A low ordecreasing arm circumference for thesepopulations would signal the loss ofmuscle mass, a serious sign, possiblyindicative of protein-energy malnutritionor starvation. MUAC is usually unaffectedby edema common in famine, and is asensitive reflection of tissue loss and isindependent of height.

The use of MUAC for emergencyprogram screening has limitations. Thechoice of cutoffs is challenging as there isa lack of an understanding of thefunctional significance of different levels.



78

9.

Collins et al. (2000) recommends thefollowing MUAC cutoffs for screeningadult admissions to feeding centers.

Table A4.2. MUAC cutoffs for Screening Moderate and Acute Adult Undernutrition

Level of undernutrition MUAC (cm)

Moderate <18.5

Severe <16.0

MUAC is independent of pregnancy orlactation status and therefore can be usedas an effective indicator of women’snutritional status throughout thereproductive years. MUAC is more usefulthan weight during pregnancy, as it varieslittle during pregnancy. Oneconsideration to take when using MUACto determine women’s nutritional status isthe age structure of the communitybecause MUAC increases with maternalage. Cut-offs for MUAC can fluctuatebetween ethnic groups, therefore localreferences may need to be established. Itis suggested to report change in meanMUAC over time rather than use a ill-defined cut-off.

4.4 Skinfold thicknessSkinfold thickness and arm circumferenceare two measurements that indirectlyassess two important components of abody: fat and fat-free mass. The reasonthat measuring these components isimportant is that fat is the main storageform of energy and fat-free mass, usuallymuscle, is a good indicator of the proteinreserves of a body. Skinfold thickness measures fat locatedjust underneath the skin (subcutaneousfat), which is a proxy indicator of totalbody fat. Measurements can be done in anumber of sites, including: the triceps (theback of the upper arm); the biceps (thefront of the upper arm); and thesubscapular area (the site just below theshoulder blade). The tool that measuresskinfold thickness is called skinfoldthickness calipers. Use of skinfoldthickness calipers requires specializedtraining and supervision to ensureaccurate and precise measurements. It is

not recommended for use in Title IImonitoring and evaluation.

Elderly Anthropometric AssessmentElderly is defined by WHO as those 60years old and over. Height measurementin older people can be problematic.Accurate measurements may not bepossible if the person cannot standcompletely erect.

Assessment techniques and cut-offs forelderly do not differ from those used foradult assessment. Assessing response toan intervention is possible by comparingchange in BMI over time. As with adults,reference standards can be developedlocally. All elderly assessments shouldnote the difficulty and lack of precisioninherent in elderly anthropometrics.(Ismail, Manandhar 1999)


9.

79

5.AdolescentAnthropometric

Indicators

AP

PE

ND

IX

Adolescents are defined by WHO as thosein the age range of 10 to 19 years. Thenutritional status of this age group isdifficult to assess because there is not areference standard for adolescents andthere is a growth spurt which occurs withpuberty which occurs at different ages.This limits the ability to use a referencestandard even if it is developed locally.The ACC/SCN has produced a paper onassessing nutrition status for emergency-affected populations (Woodruff, andDuffield. Adolescents: Assessment ofNutritional Status in Emergency-Affectedpopulations. SCN Geneva, July, 2000).acc.unsystem.org/scn

The ACC/SCN recommends that due toa lack of validated anthropometricprocedures for adolescents, anthro-pometry is not used without examiningother population sub-groups and otherdeterminants of nutrition and foodsecurity.

As with children, adolescentanthropometric assessment is used toreflect undernutrition. Anthropometry isalso used to reflect over nutrition but thisis not the focus of this guide. Under-nutrition in adolescents is characterizedby patterns of acute and chronicdeficiency of energy, protein andmicronutrients including vitamins andminerals. Often a person is affected byboth acute and chronic deficiency in all orsome of the key nutrients. Themanifestation of the deficiency and the

measurement is, complicated to deter-mine and the functional significanceunclear. Undernutrition is characterizedby a lack of food and while specificnutrient deficiencies occur, such aspellagra due to a lack of niacin, theprimary cause is more general. We arelearning more about specific nutrientrequirements for diseases such asHIV/AIDS but the ability of anthro-pometrics to identify these conditions islimited.

As with adults, adolescent anthropometricassessment is used for several purposesincluding:

• screening or targeting individuals forsome sort of intervention or action such assupplementary feeding during faminerelief,

• surveillance or monitoring of changesin prevalence and coverage in groups orpopulations to trigger a response includinggraduating from an intervention , and

• evaluating the impact of activities orinterventions.

Anthropometry is used to describe thenutritional situation in a population andthis can be useful for problem analysisand for evaluation. Because thedeterminants of nutrition are so many, it isimportant to examine other factors thanjust anthropometry such as the foodsecurity situation, levels of illness, caregiving practices and so on.



80

9.

There are a number of recommendationsfor different purposes. The reader iscautioned that some of the recom-mendations (including WHO) are beingquestioned (see Woodruff and Duffield,2000). The complex physiologicalchanges, pubertal development, inter-ethnic differences in genetic growthpotential, and the different determinantsof body size and shape makes rigidrecommendations difficult.

The recommendations for emergencyscreening should be for indicators that areeasy to measure and do not requirecumbersome or complicated equipmentand procedures. The indicator should bedetermined taking into account differencesbetween the survey and referencepopulations in age, sexual developmentand ethnicity. Most commonly foundmeasurements in emergency situationsare weight, height and mid-upper-arm-circumference (MUAC). Measuring heightcan be challenging especially inemergency situations and for severelymalnourished people who are feeble anddisabled, height measurement may beimpossible. The presence of famineedema is a serious indicator of nutritionalstress and the accumulation of fluiddistorts weight measures. Any situation ofedema would be an indication for a rangeof nutrition interventions.

For targeting interventions andassessing the situation, thinness measuresare recommended. Thinness can bereflected by percent median weight-for-height or BMI-for-age and the RohrerIndex. Thinness is especially of concernamong those adolescents who have not yetfinished their growth spurt. Pregnancyadds weight to the girl and will distort thevarious weight based measures. Duringpregnancy, measures of weight changeand MUAC are recommended.

For assessing response to anintervention for adolescents, BMI shouldbe used for programs designed to reducethe prevalence of thinness. BMI can becompared to local reference data orchanges from pre- and post-interventioncan be compared.

It should be stressed that there is a lackof data to relate the specific indicator and

its cutoff with health or survivaloutcomes. This means that to defineundernutrition in adolescents, we do nothave evidence of the choice of indicator orcutoff that exists with children under theage of five years. Weight for height andBMI indicators need to be examined basedon an accurate determination of age. Thismay not be possible in many situationswhere age is unknown.

The ACC/SCN recommend thatclinical criteria be used for screening fortherapeutic feeding. Surveys, theysuggest, should correct for different agesof sexual maturation if the age ofmaturation in the survey populationdiffers from the reference population.This is likely if the reference population isfrom a developed country.

The recommendation for screening inPre-Pubertal Adolescents, is to use weightfor height as the index of choice (usingweight for height reference standards).

For Post-Pubertal Adolescents, BMIshould be used and compared with theinternational reference standards.

Height-for-age. The measure for heightfor estimating stunting during adol-escence is the same as it is for youngchildren. Stunting reflects chronic mal-nutrition. The height is compared to theheight of adolescents of the same sex andage in the NCHS reference population.Growth charts for the US are available at:www.cdc.gov/growthcharts/. The cut offof <-2 Z-scores is also used. This measureis limited because height varies muchmore among healthy adolescents than itdoes for preadolescent children, making itdifficult to establish reliable benchmarks.Locally defined cut-offs can providegreater accuracy.

Weight-for-height is problematicbecause at a given height, the medianweight differs depending on age. Thisdoes not allow analysis of weight-for-height across wide age categories.

BMI is the foundation of accurateanthropometric assessment for adol-escents. However, without age, BMI datais quite limited for adolescents.


9.

81

BMI-for-age. Median and less thanmedian BMI-for-age varies little amongwell nourished populations. Highpercentage BMI at any age has shownvariations and is a less accurate indicatorfor overweight assessment. BMI-for-ageis also inaccurate for stunted individuals.Cut-off values are not well tested forassessing risk and response tointerventions. Local reference standardsshould be developed. Growth charts forthe US are available atwww.cdc.gov/growthcharts/ but theACC/SCN cautions on the use of thesereference standards. It should be notedthat BMI-for-age is not a straightforwardconcept and has not been examined for itsability to predict outcomes amongmalnourished adolescents.

Rohrer Index (weight/height3) iscalculated as the weight in kilogramsdivided by height in meters cubed. Anadolescent girl weight 24.4 kilograms andis 132.3 cm tall would have a RohrerIndex of 10.5. There is some evidence thatthe Rohrer Index is less age dependentand can be used like BMI. There are noreference standards for the use of thisindex.

Percent median weight-for-height. Inthe absence of strong, simple adolescentindicators, recommendations are to usepercent of the median weight-for-height.Using weight-for-height has the limitationbeing an indicator of current or acutemalnutrition. Because of its response toshort-term influences, wasting is not usedto evaluate Title II programs but may beused for screening or targeting purposesand is sometimes used for annualreporting.

Mid-upper arm circumference (MUAC)is the circumference of the left upper armmeasured in centimeters. The point isbetween the tip of the shoulder and theelbow. The use of MUAC and theequipment for the measurement isdetailed in section 5 of this Guide. Armcircumference is measured with specialcircumference measuring tapes orcircumference insertion tapes.

Because the equipment is lightweight andtraining to do MUAC is straightforward,MUAC is utilized for screening inemergency situations when nutritionalstatus information is needed immediatelyfor large groups of people, especiallychildren. The indicator is useful forassessing acute adult undernutrition andfor assessing the prevalence ofundernutrition at the population level.

While arm circumference measuresboth muscle and fat, some populationswould be expected to have very littlesubcutaneous fat on their arms. A low ordecreasing arm circumference for thesepopulations would signal the loss ofmuscle mass, a serious sign, possiblyindicative of protein-energy malnutritionor starvation. MUAC is usually unaffectedby edema common in famine and is asensitive reflection of tissue loss and isindependent of height.

The use of MUAC for emergencyprogram screening has limitations. Thechoice of cutoffs is challenging as there isa lack of an understanding of thefunctional significance of different levels.No reference standards exist for MUAC.Careful training is required to ensure thatthe correct location is identified on thearm. This is especially important withrapidly growing adolescents. It issuggested that MUAC be presented by agecategory and sex.

There are no recommended MUACcutoffs for determining adolescentundernutrition or for admissions tofeeding centers.



82

9.

6. Standardization ofAnthropometricMeasurementsA

PP

EN

DI

X

The training of personnel on specificmeasurement and recording techniquesincludes not only theoretical explanationsand demonstrations, but also anopportunity to allow participants topractice the measurement techniques, aswell as reading and recording the results.This practice is more efficient when alarge number of children are available.

Once all personnel have adequatelypracticed the measurement and recordingtechniques, and feel comfortable withtheir performance, standardizationexercises can be carried out. Eachexercise is performed with a group of 10children whose ages fall within the pre-established range for the study. Asequential identification number isassigned to both children and staff. Toconduct the exercises the following areneeded:

• balances/scales and height boards;

• pens; and

• sufficient Anthropometric Standard-ization Forms 1 and 2, to record theexercise number, name and number of themeasurer, date on which the exercise isconducted, and a sequential listing ofchildren with their name, age andidentification number.


9.

83

Form 1. Formula for Anthropometric Standardization

Exercise number:

Name of measurer:

Measurer's code:

Date / / 2000

Name

Number of large differences

Number of medium differences

Number of small differences

=

(0.3 Kg or more)

(0.2 Kg)

(0.0 or 0.1 Kg)

Whole circles

Empty circles

Without circles

1

2

3

Age in months

No.

1

2

3

4

5

6

7

8

9

10

My measure Standardmeasure

Differencesign ( + , -)

Weight



84

9.

Form 2. Formula for Anthropometric Standardization

Exercise number:

Name of measurer:

Measurer's code:

Date / / 2000

Name

Number of large differences

Number of medium differences

Number of small differences

No. DIFS. Sign (+): No. DIFS. Sign (-):

(1.0cm or more)

(0.6-0.9cm)

(0-0.5cm)

Whole circles

Empty circles

Without circles

1

2

3

Age in months

No.

1

2

3

4

5

6

7

8

9

10

My measure Standardmeasure

Differencesign ( + , -)

Height

=


9.

85

Measurement and RecordingBefore carrying out the exercise, thesupervisor carefully weighs and measureseach child and records the results withoutany of the trainees seeing the results. Foreach exercise, a group of up to 10measurers will conduct the measure-ments in a pre-determined order. Eachchild will remain at a fixed location. Thedistance between each child should be bigenough to prevent measurers seeing/hearing each others results.

At the beginning of an exercise, eachmeasurer and assistant is paired with achild. Once the children and themeasurers have been positioned with theirrespective materials and instruments, thesupervisors should instruct the measurersto begin the measurements following thepre-established sequence. The measurercarefully conducts the measurements andclearly records in ink the results on theanthropometric standardization form (MYMEASURE column) next to the child’sidentification number. The measurersremain with the child until the supervisorinstructs them to move. Once results arerecorded, corrections are not allowed.When all the measurers have conductedtheir measurements, the supervisorshould instruct them to move to next childfollowing the numerical order andrequests that they wait for instructions tobegin the measurement. This process isrepeated until all children have beenweighed and measured by all themeasurers.

Use the same equipment to measureeach child’s weight and stature.Measurers and assistants should rotate toconduct the measurement, but theequipment remains stationed next to eachchild. Only one pair of measurers shouldbe with a child at any one time. Talkingbetween measurer-pairs during thisexercise is not allowed.

The supervisor should take advantageof the standardization exercises tosystematically observe each measurer’sperformance using the MeasurementTechniques Observation Form 3. Thisform contains a list of the most importantsteps of each measurement technique, thatallows the supervisor to record if each step

was completed appropriately, and to laterdiscuss the results of these observationswith the staff.



86

9.

Form 3. Measurement techniques observation

1Observer #

Exercises 1 2 1 2 1 2 1 2

Position of equipment

Adjustment to zero

Clothes

Child's attitude

1

2

3

4

Child's position

Reading time

Reading angle

Reading

5

6

7

8

Value

Observer

Supervisor

Difference

9

10

11

12

13

14

2 3 4

Weight

VA

L

U

E


9.

87

1Observer #

Exercises 1 2 1 2 1 2 1 2

Position of equipment

Adjustment to zero

Clothes

Child's attitude

1

2

3

4

Child's position

Reading time

Reading angle

Reading

5

6

7

8

Value

Observer

Supervisor

Difference

9

10

11

12

13

14

2 3 4

Height

VA

L

U

E



88

9.

Height

Form 4. Example of overplacing of forms to compare measurers

Standardization ofAnthropometrics Weight

Large differences (1.0 or more) = whole circles

Medium differences (0.6 or 0.9cm) = empty circles

Small differences (0.5 or less) = no circles

# of diffs. sign + = ( )LESS # of diffs. sign -

Total

Total

Total

Total

Total

Total

Total

Total

Total

1

2

3

Name of measurer

Exercise #1

DifferenceA - B

sign (+ , -)

StandardMeasurement

B

Measurement

A

Age inmonths

#

1

2

3

4

5

6

7

8

9

10

DifferenceA - B

sign (+ , -)

DifferenceA - B

sign (+ , -)

DifferenceA - B

sign (+ , -)

Exercise #1 Exercise #1 Exercise #1

Measurer A Measurer B Measurer C Measurer D

( ) ( ) ( )

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) . ( ) . ( ) .

( ) .

( ) .

( ) .

( ) .

( ) .

( ) .

( ) .

( ) .

( ) .

( ) .

( ) ...

..

..

..

..

..

..

..

..

..

( ) . ( ) .


9.

89

Form 5. Tracking measurer's progress

Name of measurer-pair

No.

Name of supervisor

Name of supervisor

Result exercise No. #

Date

Number of full circles

Number of empty circles

Number without circles

Total markings

Name of measurer-pair

No.

Name of supervisor

Name of supervisor

Result exercise No. #

Date

Number of full circles

Number of empty circles

Number without circles

Total markings

Height Weight

Height Weight



90

9.

AnalysisOnce all the children have been measuredby all the measurer-pairs, the supervisorshould meet with the group to analyze theresults of the exercise and the measurersread out-loud the results of his/hermeasurements for each child. Themeasurers should record these results intheir respective forms, under theSTANDARD MEASURE column (seeForms 1 and 2). Next, each of themeasurers should calculate the differencebetween MY MEASURE and STANDARDMEASURE for each measurement andchild, and record the result on the sameform under the DIFFERENCE column,using the corresponding + or - sign: if themeasurer’s measurement is larger thanthe supervisor’s measurement (standardmeasurement), the sign is positive; if themeasurer’s measurement is lower thanthe supervisor’s measurement, the sign isnegative. Following this procedure, themeasurers should draw a circle to theright of the large and medium differencesas follows:

Measurement Whole Circle No Circle

Empty Circle

Child’s weight 0.2 kg < 0.2 kg

0.3 kg or more

Child’s length 0.6 to 0.9 cm < 0.6 cm

/height 1.0 cm

or more

Each measurer then totals the number oflarge differences (marked with a wholecircle) and the number of mediumdifferences (marked with an empty circle),and records the totals in the correspondingboxes on the lower part of the form (box 1for large differences and box 2 for mediumdifferences). The sum of all smalldifferences or no differences should becalculated and recorded in box 3.

Finally, the difference between thenumber of positive and negative differences(excluding zeros) should be calculated andrecorded in the big box on the form with itscorresponding sign. For example, if thereare 6 positive differences, 3 negativedifferences, and a zero, the result is +3; if

there are 8 negative differences and 2positive differences, the result is -6.

InterpretationThe interpretation of the standardizationexercise results is made by the measurerswith the supervisor’s help. The purpose isto detect differences, identify theirpossible causes, and correct them. Toachieve this, it is important to take intoaccount the size of the differencesbetween each measurer and thesupervisor, as well as the positive ornegative sign of the differences.

1. Size of the differencesThe total number of the differences,according to their size, has already beenrecorded by line 1, 2, and 3 located in atthe bottom of the form, as follows: largedifferences in line 1, medium ones in line2, and small differences (including theabsence of differences) in line 3. As thenumber of differences in lines 1 and 2decreases, especially in line 1, theagreement between the measurer and thesupervisor increases; that is, the better thestandardization of the measurer with thesupervisor. Large differences (line 1)generally indicate carelessness in thereading or recording, or serious problemsin the measurement technique. Moderatedifferences usually indicate problems inthe measurement technique. The ideal isto obtain the largest number in line 3,(small differences or completeagreement).

In cases where large or mediumdifferences are found, the respectivemeasurer with the assistance of thesupervisor should carefully repeat themeasurement in order to identify andcorrect the cause of the differences.

2. Sign of the differencesIf the total registered in the big box is +6or more, the measurer’s measurementsare consistently larger than thesupervisor’s. For weight measurements,the most frequent causes for differencesare: not adjusting the scale to zero at thebeginning; reading the scale in an obliqueposition and not facing the scale; orreading the scale by following the


9.

91

incorrect direction. In height or lengthmeasurements, the most frequent causesfor differences are: inadequate position ofthe head or feet; a reading done in anoblique position and not facing the readingpoint of the measuring board or height-measuring apparatus; or a readingconducted by following the incorrectdirection of the scale.

If the total in the big box is -6 or less(for example, -8), the measurer’smeasurements are consistently smallerthan the supervisor’s measurements.With weight measurements, the mostfrequent causes for differences are similarto those described above for consistentlylarge measurements. With height orlength measurements the most frequentcauses are similar to those describedabove for consistently largemeasurements, in addition to flexing of thechild’s legs during measurement.

In all cases in which large or mediumdifferences are found, the respectivemeasurer with the assistance of thesupervisor should carefully repeat themeasurement in order to identify andcorrect the cause of the differences.

3. Sample exercise with 4 measurers

MeasurersA B C D

Large differences 1 2 0 0(line 1)

Medium differences 4 0 2 8(line 2)

Small differences 5 8 8 2(line 3)

Signs exercise +2 -5 +6 -6(final box)

Possible interpretation:

Measurer A Careless measurement (reading orrecording) and problems with themeasurement technique.

Measurer BCareless measurement, but no evidentproblems with the measurementtechnique.

Measurer CGenerally well done (1 to 2 moderatedifferences may be allowed), but the signtest gives +6.

Measurer DProblems are evident with themeasurement technique and the sign testgives -6.

4. Comparison between measurersThe forms used in the exercise for eachanthropometric pattern can be overlappedso that comparisons can be made betweenmeasurers (see Form 4 for example ofweight measurements). A summary tablethat shows progress made by themeasurer-pairs can also be prepared(Form 5). If the frequency, magnitude,and sign of the differences with thesupervisor are similar for 2 or moremeasurers, this can suggest that they mayhave common problems. If the differencesconcentrate in the measurements of 1 or 2particular children, all the measurers inconjunction with the supervisor shouldrepeat the measurements for thesechildren, in order to identify and correctany problems. In some cases problemscan be the result of children who move toomuch and are difficult to measure.

D. Replication of the ExercisesThe exercises should be repeated as manytimes as necessary until none of themeasurers have large differences (line 1),and a maximum of 1 or 2 mediumdifferences (ideally zero) and until thetendency to obtain larger or smallervalues than those of the supervisordisappears (less than 6 in the big box).Generally, this is accomplished after 2or 3exercises for weight and height. Theprogress made by the measurers duringthe standardization exercises can beobserved using Form 5.



92

9.

Figure A6.1 Recording Measurements


9.

93

7.Guidelinesfor Supervising

Surveys

AP

PE

ND

IX

In order to facilitate the supervision requiredfor quality control of anthropometric data,this guide describes the procedures thatsupervisors need to follow to routinelyconduct quality control on field data.

Focus and Content of SupervisionSupervision should have a focus onguidance and training and not be seen aspolicing. In other words, reinforcingtraining and motivating staff to guaranteegood quality data is essential. From aquality control perspective, supervisionincludes two basic activities:

1. Direct observation of the measure-ment techniques conducted by themeasurers. The supervisor routinelyobserves the performance of themeasurers while they weigh and measurethroughout the entire data collectionprocess. To do this, the supervisor shouldmake a list of the most important stepsthat need to be systematically observed, ora form similar to the one used during thetraining and standardization (Form 5),which contains a column to recordwhether each step was correctly followed.

2. Replication of measurements in 10percent of the sample. The supervisorshould repeat the measurementsconducted by the measures in 10 percentof the children (one out of every ten),preferably on different days and in arandom fashion so that the measurers will

not know which children’s measurementswill be repeated. The supervisor recordsthe results of these repeated measure-ments on the regular data collectionforms, compares these results with theones conducted by the respectivemeasurer and, if discrepancies are found,discusses the results with the measurer soas to identify the causes and correct them.

Through observation of the measurersand a careful review of results of thereplications, the supervisor can reinforcemeasurer training and correct any faults.The standardization exercises should berepeated whenever needed.

Quality Control Through VisualInspection of the FormsThe supervisor will, on a daily basis andin a systematic way, visually inspect theforms where the data are recorded. Thepurpose of this inspection is to detectmissing data, inconsistencies, recordingerrors and values outside the pre-established permissible ranges. Duringthe review, emphasis should be given tothe following points: 1. Date of measurement2. Correct identification of the form with

the mother or infant’s identification number, as well as the correct sex of the child

3. Date of birth of the child4. Weight of the child5. Length/height of the child



94

9.

Points 1 to 3 are especially important todetect errors in identification. These errorsshould be immediately corrected. Theanthropometric data (4 and 5) arereviewed with the goal of detecting errorsin recording (coding) and values outsidethe preestablished range, as follows:

• for the child’s weight, values fallingoutside the following ranges according to age:

Age (months) Range of Weights (Kg)

6 - 11 4.00-13.00

12 -17 7.00-15.00

18-23 8.00-16.50

24-29 8.50-17.50

30-35 9.00-19.00

36-41 9.50-22.15

42-47 10.00-23.00

48-53 11.00-24.50

54-59 11.50-26.00

60-65 12.00-27.50

66-71 13.00-29.50

72-77 13.50-31.50

78-83 14.00-33.50

Age (months) Range of

Length/heights (cm)

6 - 11 60-90

12 -17 65-95

18-23 70-100

24-29 75-105

30-35 80-110

36-41 80-115

42-47 85-115

48-53 90-120

54-59 90-125

60-65 95-130

66-71 95-130

72-77 100-135

78-83 100-140

Values outside the ranges should becarefully reviewed by both the supervisorand the measurers with the aim ofassuring that there were no errors inmeasurement or recording. In cases ofdoubt, the measurement should berepeated; if this is not possible the datashould be erased and replaced with thecode “999” for the weight or “9999” forheight. If it is confirmed that the data arecorrect, they are kept even though theymay be outside the range.

Once the coding has been visuallyinspected, data are entered into thecomputer. Data processing procedures forquality control will also be applied todetect possible coding errors,inconsistencies, and data outside thespecified ranges. It is important toemphasize, however, that quality controlin the field through supervision, includingthe daily and routine inspection of forms,is the only efficient procedure to detectand correct errors, omissions, andinconsistencies in a timely manner. Thefeasibility of correcting errors during dataprocessing is much lower given thatgenerally it is too late to return to theprimary source of the data, the majority oferrors detected at that late time cannot becorrected and many of them end up asmissing data.


9.

95

8.Title IIGeneric Indicators

AP

PE

ND

IX

The Title II Generic Indicators

Indicator

% stunted children 24-59 months (height/age Z-score)

% underweight children by age group (weight/age Z-score)

% infants breastfed w/in 8 hours of birth

% infants under 6 months breastfed only

% infants 6-10 months fed complementary foods

% infants continuously fed during diarrhea

% infants fed extra food for 2 weeks after diarrhea

% eligible children in growth monitoring/promotion

% children immunized for measles at 12 months

% of communities with community health organization

% children in growth promotion program gaining weight in past 3

months (by gender)

% infants with diarrhea in last two weeks

liters of household water use per person

% population with proper hand washing behavior

% households with access to adequate sanitation (also annual

monitoring)

% households with year-round access to safe water

% water/sanitation facilities maintained by community

% households consuming minimum daily food requirements

number of meals/snacks eaten per day

number of different food/food groups eaten

Level

Impact

Annualmonitoring

Impact

Annualmonitoring

Impact

Category

Health, nutritionand MCH

Water and Sanitation

Household foodconsumption

continued overleaf



96

9.

The Title II Generic Indicators - continued

Indicator

annual yield of targeted crops

yield gaps (actual vs. potential)

yield variability under varying conditions

value of agricultural production per vulnerable household

months of household grain provisions

% of crops lost to pests or environment

annual yield of targeted crops

number of hectares in which improved practices adopted

number of storage facilities built and used

imputed soil erosion

imputed soil fertility

yields or yield variability (also annual monitoring)

number of hectares in which NRM practices used

seedling/ sapling survival rate

agriculture input price margins between areas

availability of key agriculture inputs

staple food transport costs by seasons

volume of agriculture produce transported by households to markets

volume of vehicle traffic by vehicle type

kilometers of farm to market roads rehabilitated

selected annual measurements of the impact indicators

Level

Impact

Annualmonitoring

Impact

Annualmonitoring

Impact

Category

Agriculturalproductivity

Natural resourcemanagement

FFW / CFW roads


FOOD

NUTR

TECH

ASSIST

Food and Nutrition Technical Assistance ProjectAcademy for Educational Development

1825 Connecticut Ave, NW,Washington D.C., 20009-5721

Tel: 202-884 8000Fax: 202-884 8432

E-mail: [email protected]: www.fantaproject.org

Date post:	27-Oct-2021
Category:	Documents
Upload:	others
View:	12 times
Download:	0 times

Anthropometric Indicators Measurement Guide

Documents