maternal social capital and child health in vietnam

W O R K I N G P A P E R N O . 3 0

Maternal Social Capital and Child Health in VietnamTran TuanTrudy HarphamMary J De SilvaNguyen Thu HuongBill TodPham Thi LanTran Duc ThachSavitri Abeyasekera

In Vietnam there is growing concern about the potential social impact of rapid economic changes. The extent and type of social connectedness, or social capital, may be changing. Studies from other developing countries have demonstrated that social capital is often independently associated with various indicators of well-being, including some aspects of human capital (health and education status). In Vietnam there has only been one previous quantitative study of social capital and this did not consider associations with well-being. The Young Lives project in Vietnam allows the examination of the relationship between maternal social capital and child well-being. With a sample of 1,953 mothers of one-year-olds and 954 mothers of eight-year-olds across five provinces, this study examines whether maternal social capital is associated with child health.

This study found low levels of structural social capital (associational life) and citizenship and high levels of cognitive social capital (trust, etc) and support. While poorer women have even lower levels of structural social capital than their better-off peers, patterns across rural and urban areas are similar.

Maternal social support and high cognitive social capital show the most consistent associations with child health, while associations between maternal social capital and child health are stronger among one-year-old children than among eight-year-olds, with the exception of child mental health and risk of life-threatening illness. The most consistent associations were seen with the mental health of eight-year-olds, with all four indicators of social capital being significantly associated with mental health after other variables have been controlled for. Lastly, there was some evidence to suggest that active membership of formal organisations in Vietnam may be damaging to the health of eight-year-olds but not one-year-olds, with active participation in formal groups being significantly associated with an increase in stunting among eight-year-olds.

The main components of social capital associated with child health are social support and cognitive social capital. High levels of maternal social support are positively correlated with having an educated partner, living in an urban area and better maternal mental health. High cognitive social capital is positively correlated with living in a rural area, having an occupation other than agriculture and better maternal mental health.

There is a need to focus on maintaining the high levels of cognitive social capital and raising the current low level of maternal structural social capital in Vietnam.

Key words: social capital, child health, Vietnam, quantitative

Published by

Young Lives Save the Children UK 1 St John's Lane London EC1M 4AR

Tel: 44 (0) 20 7012 6796 Fax: 44 (0) 20 7012 6963 Web: www.younglives.org.uk

ISBN 1-904427-31-6 First Published: 2006

All rights reserved. This publication is copyright, but may be reproduced by any method without fee or prior permission for teaching purposes, though not for resale, providing the usual acknowledgement of source is recognised in terms of citation. For copying in other circumstances, prior written permission must be obtained from the publisher and a fee may be payable.

Designed and typeset by Copyprint UK Limited

Young Lives is an international longitudinal study of childhood poverty, taking place in Ethiopia, India, Peru and Vietnam, and funded by DFID. The project aims to improve our understanding of the causes and consequences of childhood poverty in the developing world by following the lives of a group of 8,000 children and their families over a 15-year period. Through the involvement of academic, government and NGO partners in the aforementioned countries, South Africa and the UK, the Young Lives project will highlight ways in which policy can be improved to more effectively tackle child poverty.

The Young Lives Partners Centre for Economic and Social Studies (CESS), India

Department of Economics, University of Addis Ababa, Ethiopia

Ethiopian Development Research Institute, Addis Ababa, Ethiopia

General Statistical Office, Government of Vietnam

Grupo de Análisis para el Desarrollo (GRADE), Peru

Institute of Development Studies, University of Sussex, UK

Instituto de Investigación Nutricional (IIN), Peru

London School of Hygiene and Tropical Medicine, UK

London South Bank University, UK

Medical Research Council of South Africa

RAU University, Johannesburg, South Africa

Research and Training Centre for Community Development, Vietnam

Save the Children UK

Statistical Services Centre, University of Reading, UK

Maternal Social capital and child health in VietnaM

Maternal Social capital and child health in Vietnamtran tuantrudy harphamMary J de Silvanguyen thu huongBill todpham thi lantran duc thachSavitri abeyasekera

WorKinG paper no.30


ii

ii

prefaceThispaperisoneofaseriesofworkingpaperspublishedbytheYoungLivesproject,aninnovativelongitudinalstudyofchildhoodpovertyinEthiopia,India(AndhraPradeshstate),PeruandVietnam.Between2002and2015,some2,000childrenineachcountryarebeingtrackedandsurveyedat3-4yearintervalsfromwhentheyare1until14yearsofage.Also,1,000olderchildrenineachcountryarebeingfollowedfromwhentheyareaged8years.

YoungLivesisajointresearchandpolicyinitiativeco-ordinatedbyanacademicconsortium(composedoftheUniversityofReading,theLondonSchoolofHygieneandTropicalMedicine,LondonSouthBankUniversityandtheSouthAfricanMedicalResearchCouncil)andSavetheChildrenUK,incorporatingbothinterdisciplinaryandNorth-Southcollaboration.

YoungLivesseeksto:

Producelong-termdataonchildrenandpovertyinthefourresearchcountries

Drawonthisdatatodevelopanuancedandcomparativeunderstandingofchildhoodpovertydynamicstoinformnationalpolicyagendas

Traceassociationsbetweenkeymacropolicytrendsandchildoutcomesandusethesefindingsasabasistoadvocateforpolicychoicesatmacroandmesolevelsthatfacilitatethereductionofchildhoodpoverty

Activelyengagewithongoingworkonpovertyalleviationandreduction,involvingstakeholderswhomayuseorbeimpactedbytheresearchthroughouttheresearchdesign,datacollectionandanalyses,anddisseminationstages

Fosterpublicconcernabout,andencouragepoliticalmotivationtoacton,childhoodpovertyissuesthroughitsadvocacyandmediaworkatbothnationalandinternationallevels.

Initsfirstphase,YoungLiveshasinvestigatedthreekeystorylines–theeffectsonchildwell-beingof(i)accesstoanduseofservices,(ii)socialcapital,and(iii)householdlivelihoods.Thisworkingpaperisoneofaserieswhichconsideranaspectofeachofthesestorylinesineachcountry.Asaworkingpaper,itrepresentsworkinprogressandtheauthorswelcomecommentsfromreaderstocontributetofurtherdevelopmentoftheseideas:youcancontactYoungLivesviathewebsitewww.younglives.org.uk

TheprojectreceivedfinancialsupportfromtheUKDepartmentforInternationalDevelopmentandthisisgratefullyacknowledged.

Forfurtherinformationandtodownloadallourpublications,visitwww.younglives.org.uk

•

•

•

•

•


iiiiii

abstractInVietnamthereisgrowingconcernaboutthepotentialsocialimpactofrapideconomicchanges.Theextentandtypeofsocialconnectedness,orsocialcapital,maybechanging.Studiesfromotherdevelopingcountrieshavedemonstratedthatsocialcapitalisoftenindependentlyassociatedwithvariousindicatorsofwell-being,includingsomeaspectsofhumancapital(healthandeducationstatus).InVietnamtherehasonlybeenonepreviousquantitativestudyofsocialcapitalandthisdidnotconsiderassociationswithwell-being.TheYoungLivesprojectinVietnamallowstheexaminationoftherelationshipbetweenmaternalsocialcapitalandchildwell-being.Withasampleof1,953mothersofone-year-oldsand954mothersofeight-year-oldsacrossfiveprovinces,thisstudyexamineswhethermaternalsocialcapitalisassociatedwithchildhealth.

Thisstudyfoundlowlevelsofstructuralsocialcapital(associationallife)andcitizenshipandhighlevelsofcognitivesocialcapital(trust,etc)andsupport.Whilepoorerwomenhaveevenlowerlevelsofstructuralsocialcapitalthantheirbetter-offpeers,patternsacrossruralandurbanareasaresimilar.

Maternalsocialsupportandhighcognitivesocialcapitalshowthemostconsistentassociationswithchildhealth,whileassociationsbetweenmaternalsocialcapitalandchildhealtharestrongeramongone-year-oldchildrenthanamongeight-year-olds,withtheexceptionofchildmentalhealthandriskoflife-threateningillness.Themostconsistentassociationswereseenwiththementalhealthofeight-year-olds,withallfourindicatorsofsocialcapitalbeingsignificantlyassociatedwithmentalhealthafterothervariableshavebeencontrolledfor.Lastly,therewassomeevidencetosuggestthatactivemembershipofformalorganisationsinVietnammaybedamagingtothehealthofeight-year-oldsbutnotone-year-olds,withactiveparticipationinformalgroupsbeingsignificantlyassociatedwithanincreaseinstuntingamongeight-year-olds.

Themaincomponentsofsocialcapitalassociatedwithchildhealtharesocialsupportandcognitivesocialcapital.Highlevelsofmaternalsocialsupportarepositivelycorrelatedwithhavinganeducatedpartner,livinginanurbanareaandbettermaternalmentalhealth.Highcognitivesocialcapitalispositivelycorrelatedwithlivinginaruralarea,havinganoccupationotherthanagricultureandbettermaternalmentalhealth.

ThereisaneedtofocusonmaintainingthehighlevelsofcognitivesocialcapitalandraisingthecurrentlowlevelofmaternalstructuralsocialcapitalinVietnam.

Keywords:socialcapital,childhealth,Vietnam,quantitative


iv

contents

Preface ii

Abstract iii

1. Introduction 1

1.1 economic growth, inequality, urbanisation and social exclusion in Vietnam 1

1.2 this study’s approach to social capital 2

2. Methods 4

2.1 Sample 4

2.2 Measures 4

2.3 Fieldwork and ethics 5

2.4 data analysis 5

3. Results 7

3.1 descriptive results 7

3.2 Multivariate results: the relationship between social capital and child nutrition 14

3.3 Multivariate results: the relationship between social capital and child health 17

3.4 determinants of maternal social capital 17

4. Discussionandconclusions 25

References 28


1

1. introduction

1.1 Economic growth, inequality, urbanisation and social exclusion in VietnamTherearecurrentlyanumberofdifferenteconomicanddemographicfactorsthatmayaffectsocialrelationsinVietnam.Rapideconomicgrowth,growinginequalities,urbanisationandsocialexclusionarebrieflyconsideredbelow.

Rapidnationaleconomic growthisoftenaccompaniedbyaweakeningofsocialties,orconnectedness(GrootaertandvanBastelaer,2002).VietnamisundergoingaperiodofrapideconomicchangeanditsachievementsintermsofpovertyreductionhavebeenheraldedbydonorsandbytheVietnamGovernmentasoneofthegreatsuccessstoriesofeconomicdevelopment(JointDonorReport,2003;GoV,2003).AdecadeofGDPgrowthfiguresaveragingaround7percentperannumhaveunderpinnedahalvingofincome-expenditure-consumptionpovertyfrom58percentofthepopulationin1993to29percentin2002.

Socialtiesbetweenandwithincommunitiescanalsobeseveredinacontextofgrowinginequalities.ThegapbetweenthepoorestquintileofVietnamandtherichesthasbeenwideningbothintermsofincomeandexpendituresince1993.TheGinico-efficientincreasedfrom0.35in1992/93to0.37in1997/98andwasupto0.41in2002(WorldBank,2003).Thesituationoftheminorityethnicpopulationisespeciallyworrying:whilethepovertyrateforthemajoritypopulationdeclinedfrom54percentin1993to29percentin2002,thepovertyrateforminorityethniccommunitiesremainshighandonlydeclinedfrom86percentto70percentinthesameperiod.Infact,followingimprovementsintheperiod1993-1998,severepovertyamongthepoorestminorityethniccommunitiesactuallyincreasedinmanyregionsfrom1998to2002(UNCT,2003).

Changesinsocialconnectednesscanalsoresultfromdemographicphenomenasuchasurbanisation,migrationandtheageingofapopulation.UrbanisationisaparticularlyimportantphenomenoninVietnam.By2020itisestimatedthat55percentofthepopulationwillliveincitiescomparedtolessthan25percenttoday(UNCT,2003).RecentestimateshavesuggestedthatonemillionpeopleayeararemigratingtoVietnam’scities(JointDonorReport,2003).Newmigrantstothecitiescurrentlyfaceadoubledisadvantageinthelossoftheirtraditionalsocialnetworksandthediscriminatorypolicieswithregardtoaccesstobasicservices,housingandpovertyalleviationprogrammes(SavetheChildrenUK,2003andJointDonorReport,2003).

Social exclusion inVietnamistackledheadonbyformalstatesafetynetstargetedattheexcluded,poorandvulnerable(forexample,theHungerEradicationandPovertyReductionprogramme–HEPR;andProgramme135).However,questionshavebeenraisedbythedonorcommunityabouttheeffectivenessofthetargetingandthesubsequentimpactoftheseformalsafetynets.Coverageratesarelow:itisestimatedthatProgramme135reaches28percentofthepoor,whileHEPRactivitiesonlyreach6-13percentofthepoor.Furthermore,findingsontheimpactofHEPRaccessaremixed:whileimprovementshavebeenfoundinschoolenrolment,nopositivecorrelationhasbeenfoundwithhouseholdexpenditureorwithbetterhealthoutcomes(JointDonorReport,2003).Communityconsultationscarriedoutinsixprovincesin2001inordertoformulatethegovernment’spovertyreductionstrategybasedonitsinterimPovertyReductionStrategy(i-PRSP)alsofoundwidespread


2

dissatisfactionwithsocialsafetynetcoverageandimpact–particularlyintermsoffeeexemptionpoliciesinhealthandeducation(ShanksandTurk,2002).Andanotherstudyconcludedthatwithrespecttothefinancingofsocialprotectionmeasures,themajorcostofriskmanagementispaidforbyhouseholdsandcommunities,andthemainbeneficiariesofstate-providedmeasuresaretherelativelywell-offratherthantherelativelypoor(ConwayandTurk,2002).Inthiscontextofincompletecoveragebyformalsocialsafetynets,theroleofinformalsocialsafetynets–orsocialsupport–becomesparticularlyimportant.

Opportunitiesforinformal(beyondfamilial)socialsupportinVietnammaybeincreasing.Vietnam’semergingcivilsocietyshowssignsofincreasingnon-state,voluntary,not-for-profit,civicorganisationsbuiltaroundcommoninterestsandissues,seekingtoimprovethewell-beingoftheirmembersorthewell-beingofothers.Thismayhavemoretodowitheconomicreformpolicysincedoimoi1in1985thanspecificlegislationregardingcivilsocietyorganisationsduringthesameperiod(WischermannandVinh,2003;Dalton,2002andPedersen,2001).Othersarelessoptimistic.Participatorypovertyassessmentscarriedoutin2003in43communesacrossallregionsofthecountrybyavarietyofdifferentorganisations,followingacommonframework,foundthattheparticipationoflocalpeopleinplanning,decision-makingandimplementationoflocaldevelopmentplansislimited(JointDonorReport,2003).Nevertheless,agrowingcivilsocietyaddressingsocialandeconomicproblemsoffersnewanddiversifiedopportunitiesforindividualstoformnewsocialrelationships,increasesocialconnectednessanddevelopinformalsafetynets.

Thepotentialforatleastaslowingdownofprogressonpovertyreduction,togetherwithworryingsignsofincreasinginequality,newsocialtrendsandinadequateformalsafetynets,suggeststhatthisisagoodtimetodeepenthedebateonhowsocialcapitalcanplayaroleinVietnam’sdevelopment,howsocialconnectionsbetweenindividuals,householdsandcommunities,andthesupportwhichmayflowfromthoseconnections,contributestohumandevelopment,povertyreduction,andinthisparticularcase,childwell-being.

1.2 This study’s approach to social capital

SocialcapitalismostcommonlyassociatedwiththeresearchofBourdieu(1986),Coleman(1990)andPutnam(1993).‘Inattemptingtofindnon-economicfactorstoexplainthesuccessofcertaineconomicprocesses,theseauthorsdrawonconceptssuchastrust,participationincivilsocietyandsocialnetworks,allofwhichformthe“social capital”ofagivencommunity’(vanKemenade,2002p.7).SocialcapitalisoneoffiveassetsinthelivelihoodsconceptualframeworkdevelopedbyDFID(theothersbeingphysical,human,naturalandfinancialassets).However,thereisongoingdebateastowhetheritistrulyaformof‘capital’(eg,seeArrow,2000).Researchhasshownittohaveanindependenteffectonvariouseconomicandotherwelfareindicators,includingphysicalandmentalhealth(GrootaertandvanBastelaer,2002;DeSilvaet al,2005).Socialcapitalcanbethoughtofasthe‘value’ofsocialrelationships,reflectingthequalityandquantityofsocialrelationshipsinagivenpopulation,mostcommonlythosewithinacommunity.Indeed,Quibria(2003)suggeststhatsocialcapitalhasbecomeshorthandfor‘community’.

1 Doi moimarksarenovationperiodstartingfrom1989whenaseriesofadjustmentreformswereinitiatedinordertoboosteconomicgrowthandreducepovertyinVietnam.


3

Althoughsomewritersclaimthatthenegativesideofsocialcapitalisstillignored(eg,Quibria,2003),mostsocialcapitalresearchthesedaysacknowledgesthathighlevelsofsocialcapital,insomecontexts,canbeassociatedwithexclusionofoutsiders,restrictionsonindividualfreedomsandreinforcementofharmfulnorms.Another,morefundamental,criticismofafocusonsocialcapitalisthatitemphasisesthe‘social’ofdevelopmentattheexpenseofthe‘economic’.However,many,includingthecurrentauthors,welcometheadditionalattentiontonon-marketsocialinteractions.

TheapproachtosocialcapitalinYoungLivesavoidsfocussingoneithertrust, norms,etc, or interactionsbutconsidersboth,asrecommendedbymostrecenttextsonsocialcapital(eg,DasguptaandSerageldin,2000).However,thisstudyrecognisestheimportanceofseparating outstructuralsocialcapital(objectivemeasuresofwhatpeople‘do’,suchasmembershipofnetworks)fromcognitivesocialcapital(subjectivemeasuresofwhatpeople‘feel’,suchasnotionsoftrustandreciprocity)(Harpham,GrantandThomas,2002).Itisimportanttoseparatetheseindicatorsofsocialcapitalastheymaynotbecorrelated.Allmeasuresrefertothe‘community’thereforethisstudyonlyaddressesbonding(withincommunity)socialcapitalandnotbridging(linkingtoothercommunitiesorinstitutions)socialcapital.Thisstudyalsoavoidscoveringboththecausesandconsequences(oroutcomes)ofsocialcapitalinthemeasureofsocialcapitalitself.AsPortes(1998)andQuibria(2003)pointout,thisstillhappensinsomestudies.

Currentresearchviewssocialcapitalasboththepropertyofindividuals(thedirectimpactofanindividualparticipatinginanetwork),andasanecologicalcharacteristic(theindirectimpactofnetworksirrespectiveofparticipationinthosenetworks).Individualsocialcapitalismeasuredbyaskinganindividualabouttheirownparticipationinnetworks(structural),andaboutthequalityofthosenetworks,forexamplewhethertheyaretrusting(cognitive).Ecologicalsocialcapitalismeasuredbyaskingarepresentativesampleofthecommunitytheaboveissues,andaggregatingtheanswerstothecommunitylevel,providingthesamescoreofsocialcapitalforeverypersoninthecommunity.Forthepurposesofthispaper,socialcapitalisregardedasacharacteristicofindividualsandnotasanecologicalvariable.

ThisstudyisthefirstquantitativeexaminationofmaternalsocialcapitalinVietnamthatconsidersthecorrelatesofmaternalsocialcapitalanditseffectuponchildhealth.Highlevelsofmaternalsocialcapitalmaypositivelyaffectchildhealthbyenablingmotherstoaccessmoreservices(health,education,childcare),byenablingthemtoaccessmoreassets(jobs,money,durables,etc),andbybeingpositivelyrelatedtomaternalphysicaland/ormentalhealth.Whilepreviousresearchindevelopedanddevelopingcountrieshasdemonstratedpositiveassociationsbetweenadultsocialcapitalandvariousadult(orhousehold)indicatorsofwell-being,theassociationbetweenmaternalsocialcapitalandchildhealthhasnotpreviouslybeenexamined(ProductivityCommission,2003).


4

2. Methods

2.1 Sample

ThisstudyusesdatafromtheYoungLivesproject,aninternationalstudyofchildhoodpovertyinEthiopia,India,PeruandVietnam.FulldetailsofthesampleselectionfortheVietnamesepartofthisstudycanbefoundatwww.younglives.org.uk.Briefly,arandomsampleof100one-year-oldsand50eight-year-oldswastakenfromeachof20sentinelsites,over-sampledtoincludepoorareas,makingatotalof2,000one-year-oldsand1,000eight-year-olds.These20sentinelsitescomprised31communes,orcommunities.Thechild’sprimarycaregiverwasinterviewed,andforthepurposesofthisstudythesamplewasrestrictedtobiologicalmothers,resultingin1,953mothersofone-year-olds,and954mothersofeight-year-olds.

2.2 Measures

SocialcapitalwasmeasuredusingashortenedversionoftheA-SCATtooldevelopedbyHarpham,GrantandThomas(2002)–aseriesoftenyes/noquestions.ThequestionnairehasbeenqualitativelyvalidatedinVietnam. Structural social capital wasmeasuredby(1)askingrespondentsaboutactivemembershipofformal(women’sunion,co-op,tradeunion,political/social)andinformal(religious,credit,lifeinsurance,sports)groupsintheircommunityduringthelastyear;(2)byrecordingsupportreceivedfrominformal(family,relatives,neighbours,friends,religiousleaders)andformal(communeleaders,governmentofficials,politicians,NGOs)networksduringthelastyear;and(3)byaskingaboutjoiningtogetherwithothercommunitymemberstoaddressanissue/problemandcommunicationwithcommunityleaders(seeTable3).Cognitive social capital wasmeasuredbyaskingrespondentsquestionsabouttrust,senseofbelonging,communityharmonyandperceivedfairness.

The health outcomes for one-year oldsincludenutritionalstatus,physicalhealthstatusperceivedbythemotherincomparisontootherchildren,perceivedlong-termhealthproblem,whetherthechildhasbeensosickthemotherthoughttheymightdie,injuryoraccidentssincebirth,andillnessinthelast24hours(measuredusingquestionsdevelopedbyWHOforscreeningdiarrhoeaandacuterespiratoryinfection(ARI).Threeindicatorswereusedtomeasurenutritionalstatus:weight-for-agez-scores(WAZ),height-for-agez-scores(HAZ),andweight-for-heightz-scores(WHZ).AllwerecomputedusingtheNutStatmoduleofEpiInfo2002withtheWHOInternationalGrowthReferenceasastandard,2andacut-offpointof–2SDusedtoclassifycasesofmalnutrition.

The health outcomes for eight-year-oldsincludeallthoselistedabove,andmentalhealth.ChildmentalhealthwasmeasuredusingtheStrengthsandDifficultiesQuestionnaire(SDQ)withaTotalDifficultiesScorerangingfrom0to40generatedbysummingthescoresfromfourscales(emotionalsymptomsscore,conductproblemscore,hyperactivityscoreandpeerproblemscore).Ascoreof<=13wasclassifiedas‘normal’,14-16‘borderline’and>=17‘abnormal’.

Otherfactorswhichmayconfoundtheassociationbetweenmaternalsocialcapitalandchildhealth,orbedeterminantsofmaternalsocialcapital,wereexplored.Thesecomprise:maternalfactors(age,religion,ethnicity,educationlevel,occupation,numberofyearslivedincommune,andmentalhealthstatusmeasuredusingtheSRQ20withacut-offof7/8todetermineaprobablecaseofcommon

2 NutStatisanutritionanthropometryprogrammethatcalculatespercentilesandz-scoresusingeitherthe2000CDCortheWHOgrowthreference(www.cdc.gov/EpiInfo/).InanalysisofYLVietnamchildnutritionalstatus,weusedtheWHOgrowthreferenceoption,asthe1983WHOReferenceGrowthChartscontinuetobepreferredforinternationalcomparison.


5

mentaldisorder);partnerfactors(partnerathome,age,education,occupation);householdfactors(householdwealth,numberofeconomicshocksinlastthreeyears,numberofchildrenunderfiveinhousehold);childfactors(age,sex)andcommunityfactors(rural/urban).

Householdwealthismeasuredusingahouseholdwealthindex(WI)–ascorebetweenzeroandoneconstructedasanaverageofthreecomponents:(1)housingquality,whichisthesimpleaverageofroomsperperson,floor,roofandwall;(2)consumerdurables,beingthescaledsumofnineconsumerdurables(radio,bicycle,TV,electricfan,motorbike,refrigerator,landline,mobilephoneandcar/truck);and(3)servicesofdrinkingwater,electricity,toiletandfuel.Thewealthindexisdividedintofourgroups:<0.25‘poorest’,0.25-<0.5‘verypoor’,0.5-<0.75‘lesspoor’,>=0.75‘betteroff ’.

2.3 Fieldwork and ethicsEthicalapprovalwasgrantedbytheVietnameseUnionofScientificandTechnologicalAssociations,LondonSouthBankUniversity,LondonSchoolofHygieneandTropicalMedicineandReadingUniversity,UK.Datawerecollectedinlate2002,withhouseholdinterviewsperformedbytheGeneralStatisticalOffice(GSO)staffandanthropometricmeasurementstakenbytheHanoiResearchandTrainingCenterforCommunityDevelopment(RTCCD)researchteam.

2.4 Data analysisDatawereanalysedusingthe‘surveycommands’inStata8.0(StataCorp,2003).Samplingweightfactor(pweight)wasthereverseoftheprobabilityofeligiblechildrenbeingsampledineachsentinelsite.Stratawasdefinedassentinelsite,andprimarysamplingunitasthehousehold.

TheanalysisfollowedtheconceptualframeworkoutlinedinFigure1.Firstly,thedistributionofchildhealthoutcomesandmaternalsocialcapitalbyageofchild,location(rural/urban)andhouseholdwealthindexwasexploredusingPearsonchi-squaretestsforcategoricalvariablesandt-testsforcontinuousvariables,correctedforthesurveydesign.GiventhesituationinVietnam,wheretheroleofgovernmentincivilsocietylifehasbeenchangingsincethecountrymovedawayfromacentrallyplannedtowardsamarked-basedapproachtosocialdevelopment,theconceptsof‘formal’–belongingtothegovernment–and‘informal’–establishedandsustainedindependentlyfromthegovernment,groupornetwork,areusedinanalysingstructuralsocialcapital.Secondly,appropriatemultivariablemodels(eitherlinearorlogistic)foreachchildhealthoutcomewerefittedtoexploretheassociationbetweeneachspecificdimensionofmaternalsocialcapitalandchildhealthoutcomes,controllingforpotentialconfoundingfactors.Toclarifytheconceptofformal/informalgroupmembershipandsocialsupportreceived,modelswerebuiltwiththefollowingvariables:anyversusnone,anyformalversusnone,andanyinformalversusnone.

Thepredictorsofthosematernalsocialcapitalcomponentsfoundtobesignificantlyassociatedwithchildhealthoutcomesweresubsequentlyexploredinadditionalmultivariablemodels.


�

Figure 1: Conceptual framework

MaternalsocialcapitalMembershipofnetworksSocialsupportCitizenshipCognitivesocialcapital

Æ

Æ

Æ

Æ

MaternalsocialcapitalMembershipofnetworksSocialsupportCitizenshipCognitivesocialcapital

Æ

Æ

Æ

Æ

DeterminantsorconfoundersMaternalfactorsPartner’sfactorsChildfactorsCommunityfactors

Æ

Æ

Æ

Æ

DeterminantsorconfoundersMaternalfactorsPartner’sfactorsChildfactorsCommunityfactors

Æ

Æ

Æ

Æ

ChildhealthoutcomesNutritionstatusPerceivedhealthstatusLong-termhealthproblemLife-threateningillnessIllnessin24hrs/2weeksMentalhealth

Æ

Æ

Æ

Æ

Æ

Æ

ChildhealthoutcomesNutritionstatusPerceivedhealthstatusLong-termhealthproblemLife-threateningillnessIllnessin24hrs/2weeksMentalhealth

Æ

Æ

Æ

Æ

Æ

Æ


7

3. results

3.1 Descriptive results

Child health outcomesChildhealthoutcomeindicatorsarenecessarilydifferentforone-year-oldscomparedtoeight-year-olds.Theanalysesbelowwillthereforelinkchildoutcomestoeitherthesocialcapitalofthemothersofone-year-oldsorthemothersofeight-year-olds.

Figures2and3describenutritionaloutcomesbysexandlocationforeachchildagegroup.Accordingtointernationalstandards,theprevalenceofunderweightchildrenishighinbothone-andeight-year-oldgroups(from20to<30%WAZ<-2SD).Incontrast,stuntingratesarelow(<20%HAZ<-2SD)inone-year-oldsandmedium(from20to<30%)ineight-year-olds.Inbothagegroupsboysandchildrenlivinginruralareashaveworsenutritionalindicatorsthangirlsandchildrenlivinginurbanareas.

Tables1and2presentindicatorsofchildhealthbysexandlocationforone-year-oldsandeight-year-oldsrespectively.Eight-year-oldsaremorethanthreetimesmorelikelytosufferfromlong-termhealthproblemsthanone-year-olds,while21percentofeight-year-oldsareprobablecasesofmentalillness.Incontrasttothenutritionalindicators,fewsignificantdifferencesbetweenruralandurbanareas,orbetweengirlsandboysforeitheragegrouparefoundwiththephysicalhealthindicators.However,childrenofbothageslivinginruralareas,andone-year-oldboysratherthangirls,weremorelikelytohavesufferedfromalife-threateningillness.


8

Figure 2

40

30

20

10

0

18.7

2523.3

34.5

14.2

24.6

19.7

32.2

Prevalence of underweight (WAZ-scores<-2SD)in one-year-olds* and eight-year-olds** by sex and location

Very high>=30%

High<30%

Medium<20%

Low <10%

Girls Boys1-yr-olds

Girls Boys8-yr-olds

Urban Rural1-yr-olds


SEX LOCATION

Figure 3

40

30

20

10

0

12.3

17.6

21.1

33.3

7.9

17.519.7

29.8

Prevalence of stunting (HAZ-scores<-2SD)in one-year-olds* and eight-year-olds** by sex and location

High30-<40%

Medium20-<30%

Low<20%

Girls Boys1-yr-olds

Girls Boys8-yr-olds



SEX LOCATION

* overall prevalence of stunting in 1-year-olds: 15.0%, ** 8-year-olds: 27.2%


�

Table 1: Indicators of child health of one-year-olds by gender and location

Child health outcomes Total n=1953

Girl n=945

Boy n=1008

Rural n=1567

Urban n=386

Health outcomes reported by mother

% perceived health worse than others 27.8 27.3 28.2 28.0 27.3

% having long-term health problems 4.4 3.5 5.2 3.7* 6.1

% life-threatening event since birth 13.3 11.2* 15.2 14.5* 9.6

% had any injury/accident since birth 9.0 8.9 9.1 9.4 8.0

% had illness in last 24 hours 17.4 15.1* 19.5 17.2 18.0

Nutrition outcome measured on child a n=1946 n=942 n=1004 n=1562 n=384

WAZ-scores (mean & 95%CI) -1.29 -1.25 -1.33 -1.37*** -1.04

HAZ-scores (mean & 95%CI) -1.03 -0.94*** -1.1 -1.14*** -0.72

WHZ-scores (mean & 95%CI) -0.74 -0.73 -0.76 -0.76 -0.70

a anthropological measurements were missing for seven children * p<=0.05, ** p<=0.01, *** p<=0.001 based on pearson’s chi-square test for categorical, and t-test for continuous, variables

Table 2: Indicators of child health of eight-year-olds by gender and location

Child health outcomes Total n=954

Girl n=476

Boy n=478

Rural n=767

Urban n=187

Health outcomes reported by mother

% perceived health worse than others 31.5 27.9* 34.9 29.5 37.2

% long-term health problems 14.3 13.8 14.9 13.6 16.6

% life-threatening illness last 3 years 10.4 9.9 11.0 11.8* 6.6

% illness in 2 weeks prior the survey 33.2 33.2 33.2 33.7 31.5

% suspected case of mental ill health(TDS scores >=17) 21.5 20.8 22.3 19.6 27.4

Nutrition outcome measured on child a

WAZ-scores (mean) -1.55 -1.46** -1.64 -1.67*** -1.20

HAZ-scores (mean) -1.42 -1.29*** -1.54 -1.51*** -1.15

WHZ-scores (mean) -0.91 -0.91 -0.91 -1.01*** -0.62

a no missing cases * p<=0.05, ** p<=0.01, *** p<=0.001 based on pearson’s chi-square test for categorical, and t-test for continuous, variables

Maternal social capitalTheaverageageofthemothersofone-year-oldsis27years,andoftheeight-yearolds35years.Ninety-eightpercentofthemothersofeight-year-olds,and77percentofthemothersofone-year-oldshavelivedinthecommuneformorethanfiveyears(12percentoftheyoungermothershavelivedinthecommunefortwoyearsorless).


10

Table 3: Individual components of maternal social capital by child age group

Mothers of 1-year-olds

Mothers of 8-year-olds Combined

n=1953 %

n=954 %

n=2907 %

STRUCTURAl SoCIAl CApITAl

Group membership

In last year, been an active member of formal group

prof/trade union 7* 10 8

Co-op 5 6 5

Women's union 17*** 31 22

political/social group 1 2 1

In last year, been an active member of informal group

Religious group 1 2 1

Credit/life insurance group 6 7 6

Sports group 0.3 0 0.2

Social support received in the last year from

Formal networks

Commune leaders 15 17 16

politicians 0.1 0 0.1

Government officials 5 6 5

Non-government/charity officials 6* 9 7

other source 7*** 12 8

Informal networks

Family/relatives 94** 91 93

Neighbours 77* 81 78

Friends (not neighbours) 71 74 72

Religious leaders 1 2 1

Citizenship activities

Joined with other households to address a problem 30*** 50 36

Talked with commune leaders about a problem in the community 4*** 8 5

CoGNITIve SoCIAl CApITAl

In general, majority of people in commune can be trusted 83** 87 84

Majority of people in commune get along well 91* 94 92

Feel part of the commune 98** 99 98

Feel people would take advantage if they got the chance 9 7 9

* p<=0.05, ** p<=0.01, *** p<=0.001. pearson’s chi-square test is used to test for each category (dichotomous variable)


11

Table3describesthedistributionofeachaspectofmaternalsocialcapital,whileTable4presentscompositematernalsocialcapitalvariablesbychildagegroup.

Levelsofstructuralsocialcapitalarelow,particularlyamongmothersofone-year-olds.While73percentofmothersofone-year-oldsdonotparticipateinanygroup,thisfallsto59percentamongmothersofeight-year-olds(p<0.01).Inaddition,only31percentofmothersofone-year-oldsreceivesupportfrombothformalandinformalnetworks,comparedto40percentofmothersofeight-year-olds(p<0.01),andmothersofolderchildrenhavenearlytwicethelevelofcitizenship(p<0.01).Giventhecommunistpoliticalstructures,theoveralllevelofcitizenship(collectiveactionwithcommunityleaders)(37percent)issurprisinglylow.

Participationinwomen’sunionsaccountsforthemajorityofnetworkingactivitiescarriedoutbyallwomen,thoughagainparticipationisloweramongthemothersofone-year-olds(17percentcomparedto31percent,p<0.001).Whilesocialsupportfromformalnetworksislow(35percent),supportfrominformalnetworkssuchasfamily,friendsandneighboursisveryhigh(96percent,94percentand95percentrespectively).

Thefacevalidityofmeasuresofstructuralsocialcapitalappearsgoodasstatisticallysignificantdifferencesbetweenthegroupsonlyappearinquestionswhereonewouldexpectadifference.Forexample,olderwomen(mothersofeight-year-olds)aremorelikelytobeamemberofawomen’sunion(partoftheofficialCommunistParty’smassorganisations)andtohavecommunicatedwithcommuneleadersorjoinedwithothercommunememberstoaddressanissueorproblem(citizenship).Thisconcurswithexistingevidencethatshowsthatyoungergenerationsarelesslikelytoparticipateinmassorganisationsandothergroups(WischermannandVinh,2003).Itmayalsoreflectthedifferentneedsofolderandyoungerchildren–themothersofone-year-oldsprobablyhavelesstimetodevotetonetworkingactivities.

Incontrasttostructuralsocialcapital,levelsofcognitivesocialcapitalareveryhighamongbothgroups(76percentand81percent).Nevertheless,mothersofone-year-oldchildrendohavesignificantlylowerlevelsofgeneralisedtrust,senseofsocialharmonyandsenseofbelongingthanmothersofeight-year-oldchildren(p<0.01).Thiscouldbeduetotheirlowerparticipationratesoralowerdurationofresidenceinacommunity.

Thefacevalidityofthecognitivesocialcapitalquestionsisgoodasthe‘negatively’phrasedquestion‘Doyouthinkthatmostpeopleinthiscommunewouldtakeadvantageofyouifgiventheopportunity’getsalowpositiveresponseof9percent.


12

Table 4: Composite maternal social capital variables by child age group

Mothers of 1-year-olds

Mothers of 8-year-olds Combined

n=1953 %

n=954 %

n=2907 %


Group membership

None 73** 59 69

Formal groups 20** 33 24

Informal groups 4 4 4

Formal and informal groups 3 4 3

Any formal groups 23*** 37 28

Any informal groups 7*** 8 7

Any groups 27*** 41 31


None 3.4 3.6 3.5

Formal networks 0.6* 1.7 1

Informal networks 65 54 61

Formal and informal networks 31** 40 34

Any formal networks 32*** 42 35

Any informal networks 96** 94 95

Any networks 97 96 97


None 69 49 63

Some (either joined together or contacted leaders) 31** 51 37


Some (< 3 'yes') 10** 6 9

Medium (=3 'yes') 14 13 13

High (=4 'yes') 76** 81 78



13

Socialcapitaldoesnotdifferbywealthgroups(Table5)but,aswouldbeexpected,thepercentageofreceivingnosupportfromanynetworkornotbeingactivecitizensishigherinruralthaninurbanareas,whilecognitivesocialcapitalislowerinurbanthaninruralareas.

Table 5: Dimensions of maternal social capital by location and wealth (n=2907)

Urban Rural poorest very poor

less poor

Better off

n=573 n=2334 n=615 n=1102 n=922 n=268


Group membership

None 67 69 82 69 64 61

Formal groups 23 24 13 25 28 27

Informal groups 5.7 3.1 4.2 3.1 3.2 7.8

Formal and informal groups 4.3 3.1 0.7 3.3 4.6 4.3

Any formal groups 27 27 13 28 32 31

Any informal groups 6 10 5 6 8 12

Any groups 33 31 18 31 36 39


None 1.1** 4.3 4.3 3.8 3.4 0.9

Formal networks 1.3 0.8 1.3 0.8 0.7 1.7

Informal networks 53** 64 69 63 59 49

Both formal and informal networks 45 31 25 33 37 49

Any formal networks 32 46 26 33 37 50

Any informal networks 95 97 94 95 96 97

Any networks 99 96 96 96 97 99


None 51*** 67 61 65 63 55

Some (either joined together or contacted leaders) 49 33 39 35 37 45


Some (< 3 'yes') 12 7.7 8.9 9.6 8.5 6.9

Medium (=3 'yes') 13 13 13 15 13 10

High (=4 'yes') 75* 79 78 76 78 83



14

3.2 Multivariate results: the relationship between social capital and child nutritionAssociationsbetweenmaternalsocialcapitalandchildhealthoutcomesvarybyspecificoutcomesandbyspecificcomponentsofmaternalsocialcapital.Amongthefourcomponentsofmaternalsocialcapital,socialsupportandcognitivesocialcapitaldisplaythemoststatisticallysignificantassociationswithchildhealth.

Associationsbetweenmaternalsocialcapital(ie,socialsupportandcognitivesocialcapital)andchildnutritionwereapparentamongone-year-olds(Table6).Mothersofone-year-oldswhoreceivesupportfromeitherformalorinformalnetworkshavesignificantlybetterchildnutritionaloutcomesintermsofweight-for-age(β=0.25,P<0.05)andweight-for-height(β=0.19,P=0.05)aftercontrollingforanumberofotherfactors.Childrenofmotherswithhighcognitivesocialcapitalhavesignificantlyhigherweight-for-age(β=0.17,P<0.01)andheight-for-ageindices(β=0.18,P<0.01)thanthosewithlowcognitivesocialcapital.

Amongeight-year-oldsveryfewconsistentassociationswerefoundbetweenmaternalsocialcapitalandchildnutrition.Onlymembershipofformalorganisationswassignificantlyassociatedwithchildnutrition,withmembershipofformalgroupsassociatedwith worse height-for-agez-scores(β=-0.17;95%CI-0.3,-0.03).


15

Tabl

e 6:

Lin

ear

regr

essi

on e

stim

ates

of m

ater

nal s

ocia

l cap

ital

com

pone

nts

as p

redi

ctor

s of

one

-yea

r-ol

ds’ n

utri

tion

out

com

e m

easu

red

by z

-sco

re n

utri

tion

al in

dice

s (d

epen

dent

var

iabl

es)

(n=1

946)

a

Chi

ld n

utri

tiona

l out

com

e

Soci

al ca

pita

lW

eigh

t-fo

r-ag

e z-s

core

sH

eigh

t-fo

r-ag

e z-s

core

sW

eigh

t-fo

r-he

ight

z-sc

ores

Cru

de

b-co

effici

ent

[95%

CI]

Adj

uste

db b-

coeffi

cien

t [9

5%C

I]

Cru

de

b-co

effici

ent

[95%

CI]

Adj

uste

db b-

coeffi

cien

t [9

5%C

I]

Cru

de

b-co

effici

ent

[95%

CI]

Adj

uste

db b-

coeffi

cien

t [9

5%C

I]

STR

UC

TU

RA

l So

CIA

l C

ApI

TA

l

Gro

up m

embe

rshi

p

Any

vers

usn

one

Æ.0

4 [-

.06,

.15]

.02

[-.0

8, .1

2].0

7 [-

.03,

.18]

.02

[-.0

8, .1

2]-.0

2[-

.11,

.08]

.002

[-.0

9, .0

9]

Any

form

alv

snon

eÆ

.05

[-.0

6, .1

7].0

4[-

.07,

.15]

.05

[-.0

6, .1

7].0

08[-

.10,

.12]

.007

[-.0

9, .1

1].0

3[-

.07,

.13]

Any

info

rmal

vsn

one

Æ.0

04[-

.17,

.18]

-.03

[-.1

8, .1

2].1

0[-

.07,

.29]

.02

[-.1

4, .1

9]-.0

9[-

.10,

.12]

-.04

[-.1

9, .1

0]

Soci

al su

ppor

t rec

eive

d fr

om:

Any

netw

ork

vsn

one

Æ.4

1***

[.19,

.64]

.25*

[.04,

.47]

.34*

*[.0

9, .5

8].1

6[-

.09,

.41]

.25*

*[.0

5, .4

6].1

9*[-

.001

, .37

]

Any

form

aln

etw

ork

vs

none

Æ.4

4***

[.21,

.68]

.25*

[-.0

2, .4

7].3

7**

[.11,

.64]

.15

[-.1

1, .4

1].2

6*[.0

5, .4

8].1

9[-

.006

, .39

]

Any

info

rmal

vsn

one

Æ.4

1***

[.19,

.64]

.25*

[.04,

.47]

.34*

*[.0

9, .5

9].1

7[-

.08,

.42]

.26*

*[.0

4, .4

6].1

9[-

.003

, .37

]

Citi

zens

hip

Som

evs

non

eÆ

-.02

[-.1

2, .0

8].0

2[-

.07,

.12]

-.04

[-.1

4, .0

6]-.0

1[-

.10,

.08]

-.01

[-.1

0, .0

8].0

2[-

.06,

.11]

Co

GN

ITIv

e So

CIA

l C

ApI

TA

l

Med

ium

vsl

owÆ

.34

[-.1

5, .2

2].0

5[-

.12,

.22]

.09

[-.1

0, .2

7]-1

1[-

.06,

.27]

-.02

[-.1

9, .1

5]-.0

2[-

.18,

.14]

Hig

hvs

low

Æ.1

8*[.0

2, .3

3].1

7**

[.03,

.31]

.18*

*[.0

3, .3

2].1

8**

[.04,

.31]

.11

[-.0

3, .2

5].0

9[-

.04,

.23]

a se

ven

mis

sing

obs

erva

tions

b

adj

uste

d fo

r: (1

) M

othe

rs: a

ge, e

duca

tion

leve

l, le

ngth

of l

ivin

g in

com

mun

e, e

thni

c gr

oup,

occ

upat

ion,

men

tal h

ealth

sta

tus;

(2)

Hou

seho

ld: w

ealth

inde

x, s

hock

eve

nts

in la

st t

hree

yea

rs; (

3) C

omm

unity

: ru

ral/u

rban

; (4)

Chi

ld: s

ex, a

ge

* p<

=0.

05; *

* p<

=0.

01; *

** p

<=

0.00

1 (c

ompa

red

with

bas

elin

e gr

oup)


1�

Tabl

e 7:

Reg

ress

ion

esti

mat

es o

f mat

erna

l soc

ial c

apit

al c

ompo

nent

s as

pre

dict

ors

of e

ight

-yea

r-ol

ds’ n

utri

tion

out

com

e m

easu

red

by

z-sc

ore

nutr

itio

nal i

ndic

es (

depe

nden

t var

iabl

es)

(n=9

51)a

Chi

ld n

utri

tiona

l out

com

e

Soci

al ca

pita

lW

eigh

t-fo

r-ag

e z-s

core

sH

eigh

t-fo

r-ag

e z-s

core

sW

eigh

t-fo

r-he

ight

z-sc

ores

Cru

deb-

coeffi

cien

t[9

5%C

I]

Adj

uste

dbb-

coeffi

cien

t[9

5%C

I]

Cru

deb-

coeffi

cien

t[9

5%C

I]

Adj

uste

dbb-

coeffi

cien

t[9

5%C

I]

Cru

deb-

coeffi

cien

t[9

5%C

I]

Adj

uste

dbb-

coeffi

cien

t[9

5%C

I]

STR

UC

TU

RA

l So

CIA

l C

ApI

TA

l

Gro

up m

embe

rshi

p

Any

vsn

one

Æ-.0

2 [-

.13,

.10]

-.06

[-.1

7, .0

6]-.0

7 [-

.20,

.06]

-.14*

[-

.26,

-.01

].0

5 [-

.07,

.17]

.06

[-.0

6, .1

8]

Any

form

alv

snon

eÆ

-.04

[-.1

6, .0

8]-.0

7 [-

.19,

.04]

-.10

[-.2

4, .0

3]-.1

7**

[-.3

0, -.

03]

.05

[-.0

7, .1

8]-.0

5 [-

.07,

.18]

Any

info

rmal

vsn

one

Æ.1

9 [-

.06,

.45]

-.11

[-.1

3, .3

6].1

7 [-

.10,

.44]

.04

[-

.23,

.31]

.14

[-.1

3, .4

0].1

5 [-

.10,

.41]

Soci

al su

ppor

t rec

eive

d

Any

vsn

one

Æ.0

04

[-.2

9, .3

0]-.0

8 [-

.38,

.21]

.08

[-.2

7, .4

4]-.0

09

[-.3

7, .3

5]-.0

8 [-

.43,

.26]

-.14

[-.4

6, .1

8]

Any

form

alv

snon

eÆ

.01

[-.2

9, .3

1]-.1

2[-

.42,

.18]

.04

[-.3

2, .4

0]-.0

9[-

.46,

.27]

-.02

[-.3

7, .3

4]-.1

0[-

.44,

.23]

Any

info

rmal

vsn

one

Æ-.0

01[-

.29,

.29]

-.08

[-.3

8, .2

1].0

8[-

.27,

.44]

-.008

[-.3

7, .3

5]-.0

8[-

.43,

.26]

-.14

[-.4

6, .1

8]

Citi

zens

hip

Non

evs

som

eÆ

.03

[-.0

8, .1

4]-.0

2[-

.13,

.09]

-.05

[-.1

7, .0

8].0

01[-

.12,

.12]

.10

[-.0

2, .2

2]-.0

4[-

.16,

.08]

Co

GN

ITIv

e So

CIA

l C

ApI

TA

l

Med

ium

vsl

owÆ

.11

[-.1

5, .3

7].1

5[-

.10,

.39]

.11

[-.1

8, .4

0].1

4[-

.13,

.41]

.05

[-.2

6, .3

6].0

9[-

.21,

.39]

Hig

hvs

low

Æ.1

0[-

.11,

.32]

.09

[-.1

2, .3

0].1

2[-

.11,

.36]

.07

[-.1

5, .2

9].0

4[-

.22,

.30]

.08

[-.1

7, .3

3]

a tw

o m

issi

ng o

bser

vatio

ns

b a

djus

ted

for:

(1)

Mot

hers

: age

, edu

catio

n le

vel,

leng

th o

f liv

ing

in c

omm

une,

eth

nic

grou

p, o

ccup

atio

n; (

2) H

ouse

hold

: wea

lth in

dex,

sho

ck e

vent

s in

last

thr

ee y

ears

; (3)

Com

mun

ity: r

ural

/urb

an; (

4) C

hild

: se

x, a

ge, m

enta

l hea

lth s

tatu

s *

p<0.

05; *

* p<

0.01

; ***

p<

0.00

1 (c

ompa

red

with

bas

elin

e gr

oup)


17

3.3 Multivariate results: the relationship between social capital and child healthAswithnutritionaloutcomes,socialsupportandcognitivesocialcapitalshowedthemostconsistentassociationswithchildphysicalhealthoutcomesamongone-year-olds.Childrenofmotherswhoreceivedsocialsupportfromeitherformalorinformalnetworkswerestrongerthanchildrenofmotherswhohadnosocialsupport(OR=0.31,P<0.001),andmotherswhoreceivedsocialsupportfromeitherformalorinformalnetworksreportedlesslife-threateningchildillnesscomparedtothosewithlittlesupport(OR=0.52and0.47respectively;P<0.05).Childrenofmotherswithhighcognitivesocialcapitalareatalowerriskofbothacuteillnessandlong-termhealthproblemsthanthosefrommotherswithlowcognitivesocialcapital(OR=0.66,P<0.05;OR=0.52,P<0.001;andOR=0.43,P<0.01,respectively)(Table8).

Amongeight-year-olds,themostconsistentassociationsarebetweengoodchildmentalhealthandhighlevelsofmaternalsocialcapital,wherebyexceptgroupmembership,allotherindicatorsofsocialcapitalaresignificantlyassociatedwithchildmentalhealth.Childrenofmotherswithhighlevelofcognitivesocialcapitalalsohaveloweroddsofhavingsufferedalife-threateningillness(Table9).Thecausalityofthisassociationmightruneitherway:hightrustandsenseofbelongingmayenableamothertoseekcaremoreeffectively,oraseriouslyillchildmayleadamothertobemorehouseboundandtonotbeabletoestablishtrustingrelationships.

3.4 Determinants of maternal social capitalAsmaternalsocialcapitalinfluenceschildhealth,itisimportanttoknowwhatinfluencesmaternalsocialcapital.Thedeterminantsofsocialsupportandcognitivesocialcapitalamongthemothersofone-year-oldsweresubsequentlyexplored,asthesevariablesshowedthemostconsistentassociationswithchildhealth.Understandingsomethingofthedeterminantsofthesevariablesmayhelpindesigninginterventionprogrammestoimprovethem.

Predictorsofhavingsocialsupportarelivinginanurbanarea(OR=1.39and1.35respectivelyforanyformalandanyinformalsourcesofsupport;P<0.01andP<0.05),havingapartnerwhocompletedsecondaryschool(OR=3.34and2.57respectivelyforanyformalandanyinformal;P<0.05),andgoodmaternalmentalhealth(Table10).Notethatethnicity,lengthofresidenceincommune,maternaleducationstatus,maternaloccupation,householdwealthstatus,economicshocksandchild’ssexarenotassociatedwithhavingsocialsupport.

Predictorsofhighcognitivesocialcapitalaregoodmaternalmentalhealth,notworkinginagriculture,andlivinginaruralarea.Notethatmaternalage,ethnicity,lengthofresidence,maternaleducation,householdwealth,householdeconomicshocksandpartner’seducationarenotassociatedwithcognitivesocialcapital.


18

Tabl

e 8:

Log

isti

c re

gres

sion

est

imat

es o

f mat

erna

l soc

ial c

apit

al c

ompo

nent

s as

pre

dict

ors

of o

ne-y

ear-

olds

’ oth

er h

ealt

h ou

tcom

es

(n=1

953)

one

-yea

r-ol

ds’ h

ealth

out

com

es

Soci

al ca

pita

lpe

rcei

ved

heal

th w

orse

(y

es =

1)

long

-ter

m h

ealth

pro

blem

(y

es =

1)

life

-thr

eate

ning

sinc

e bir

th

(yes

= 1

)

Cru

de

odds

ratio

[9

5%C

I]

Adj

uste

d a

odds

ratio

[9

5%C

I]

Cru

de

odds

ratio

[9

5%C

I]

Adj

uste

d a

odds

ratio

[9

5%C

I]

Cru

de

odds

ratio

[9

5%C

I]

Adj

uste

d a

odds

ratio

[9

5%C

I]

STR

UC

TU

RA

l So

CIA

l C

ApI

TA

l

Gro

up m

embe

rshi

p

Any

vsn

one

Æ.8

5 [.6

7, 1

.09]

.96

[.74,

1.2

4]1.

08

[.65,

1.8

1]1.

08

[.64,

1.8

2]1.

09

[.80,

1.5

0]1.

20

[.86,

1.6

8]

Any

form

alv

snon

eÆ

.85

[.65,

1.1

1].9

7 [.7

4, 1

.28]

1.0

[.55,

1.7

1]1.

0 [.5

5, 1

.76]

1.03

[.7

4, 1

.45]

1.12

[.7

8, 1

.61]

Any

info

rmal

vsn

one

Æ.8

7[.5

6, 1

.34]

.91

[.58,

1.4

3]1.

34[.6

2, 2

.9]

1.26

[.57,

2.7

8]1.

34[.7

9, 2

.29]

1.59

[.92,

2.7

4]

Soci

al su

ppor

t rec

eive

d

Any

vsn

one

Æ.3

3***

[.20,

.55]

.32*

**[.1

8, .5

4].5

9[.1

9, 1

.80]

.43

[.14,

1.3

1].4

3**

[.24,

.78]

.47*

[.25,

.88]

Any

form

alv

snon

eÆ

.32*

**[.1

9, .5

5].3

1***

[.18,

.55]

.59

[.18,

1.9

].4

0[.1

3, 1

.30]

.46*

[.25,

.86]

.52*

[.28,

1.0

0]

Any

info

rmal

vsn

one

Æ.3

3 **

*[.2

0, .5

5].3

1***

[.18,

.54]

.60

[.19,

1.8

].4

4[.1

4, 1

.33]

.43*

*[.2

4, .7

8].4

7*[.2

5, .8

8]

Citi

zens

hip

Non

evs

som

eÆ

.93

[.74,

1.1

7].8

4[.6

6, 1

.08]

.80

[.47,

1.3

8].7

8[.4

5, 1

.36]

.99

[.73,

1.3

4].9

1[.6

5, 1

.26]

Co

GN

ITIv

e So

CIA

l C

ApI

TA

l

Med

ium

vsl

owÆ

.80

[.53,

1.2

2].8

9[.5

7, 1

.37]

.72

[.32,

1.5

9].7

1[.3

2, 1

.59]

1.11

[.64,

1.9

4]1.

12[.6

1, 2

.06]

Hig

hvs

low

Æ.6

1***

[.43,

.87]

.66*

[.46,

.95]

.42*

*[.2

2, .8

0].4

3**

[.22,

.82]

.77

[.49,

1.2

2].8

0[.4

9, 1

.31]


1�

Tabl

e 8

cont

inue

d…

one

-yea

r-ol

ds’ h

ealth

out

com

es

Soci

al ca

pita

lIn

jury

/acc

iden

t sin

ce b

irth

(yes

= 1

)Il

lnes

s in

last

24

hour

s(y

es =

1)

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d aod

ds ra

tio[9

5%C

I]

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d aod

ds ra

tio[9

5%C

I]

STR

UC

TU

RA

l So

CIA

l C

ApI

TA

l

Gro

up m

embe

rshi

p

Any

vsn

one

Æ1.

11[.7

6, 1

.61]

1.17

[.80,

1.7

1]1.

04[.7

9, 1

.38]

1.11

[.83,

1.4

8]

Any

form

alv

snon

eÆ

1.14

[.77,

1.6

8]1.

23[.8

2, 1

.85]

1.07

[.80,

1.4

4]1.

14[.8

4 , 1

.56]

Any

info

rmal

vsn

one

Æ1.

04[.5

4, 1

.98]

1.04

[.54,

1.9

9].9

6[.5

8, 1

.59]

1.04

[.62,

1.7

4]

Soci

al su

ppor

t rec

eive

d

Any

vsn

one

Æ.9

3[.4

0, 2

.13]

.92

[.38,

2.1

9].6

5[.3

6, 1

.17]

.61

[.33,

1.1

1]

Any

form

alv

snon

eÆ

.83

[.35,

2.0

].8

4[.3

4, 2

.11]

.62

[.33,

1.1

7].5

9[.3

1, 1

.12]

Any

info

rmal

vsn

one

Æ.9

3[.4

0, 2

.14]

.93

[.39,

2.2

1].6

5[.3

6, 1

.17]

.61

[.33,

1.1

1]

Citi

zens

hip

Non

evs

som

eÆ

.77

[.53,

1.1

2].7

4[.5

0, 1

.10]

.90

[.68,

1.1

9].8

8[.6

7, 1

.17]

Co

GN

ITIv

e So

CIA

l C

ApI

TA

l

Med

ium

vsl

owÆ

.71

[.38,

1.3

2].7

5[.4

0, 1

.41]

.95

[.60,

1.5

0]1.

00[.6

3, 1

.59]

Hig

hvs

low

Æ.5

8*[.3

5, .9

6].6

2[.3

7, 1

.05]

.49*

**[.3

3, .7

1].5

2***

[.35,

.77]

a a

djus

ted

for:

(1)

Mot

hers

: age

, edu

catio

n le

vel,

leng

th o

f liv

ing

in c

omm

une,

eth

nic

grou

p, o

ccup

atio

n, m

enta

l hea

lth s

tatu

s; (2

) H

ouse

hold

: wea

lth in

dex,

sho

ck e

vent

s in

last

thr

ee y

ears

; (3)

Com

mun

ity:

rura

l/urb

an; (

4) C

hild

: sex

, age

*

p<0.

05; *

* p<

0.01

; ***

p<

0.00

1 (c

ompa

red

with

bas

elin

e gr

oup)


20

Tabl

e 9:

Log

isti

c re

gres

sion

est

imat

es o

f mat

erna

l soc

ial c

apit

al c

ompo

nent

s as

pre

dict

ors

of e

ight

-yea

r-ol

ds’ o

ther

hea

lth

outc

omes

(n

=954

)

eigh

t-ye

ar-o

lds’

heal

th o

utco

mes

Soci

al ca

pita

lpe

rcei

ved

heal

th w

orse

(yes

= 1

)lo

ng-t

erm

hea

lth p

robl

ems

(yes

= 1

)li

fe-t

hrea

teni

ng il

lnes

s las

t 3 yr

s(y

es =

1)

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d a

odds

ratio

[95%

CI]

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d a

odds

ratio

[95%

CI]

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d a

odds

ratio

[95%

CI]

STR

UC

TU

RA

l So

CIA

l C

ApI

TAl

Gro

up m

embe

rshi

p

Any

vsn

one

Æ.9

3[.6

9, 1

.26]

1.06

[.77,

1.4

7]1.

13[.7

7, 1

.67]

1.17

[.77,

1.7

7]1.

06[.6

8, 1

.63]

1.32

[.83,

2.1

0]

Any

form

alv

snon

eÆ

.93

[.68,

1.2

7]1.

10[.7

8, 1

.54]

.98

[.65,

1.4

8]1.

00[.6

5, 1

.54]

1.05

[.67,

1.6

5]1.

33[.8

2, 2

.16]

Any

info

rmal

vsn

one

Æ1.

08[.6

3, 1

.88]

1.05

[.61,

1.8

1]2.

0*[1

.07,

3.7

4]1.

88[.9

5, 3

.70]

1.24

[.57,

2.6

9]1.

48[.6

5, 3

.35]

Soci

al su

ppor

t rec

eive

d

Any

vsn

one

Æ.8

6[.4

1, 1

.78]

.77

[.36,

1.6

4].8

3[.3

3, 2

.10]

.78

[.31,

1.9

6].6

6[.2

4, 1

.78]

.61

[.22,

1.6

4]

Any

form

alv

snon

eÆ

.73

[.34,

1.5

5].6

5[.3

0, 1

.42]

.87

[.34,

2.2

6].8

3[.3

1, 2

.19]

.70

[.25,

1.9

5].7

1[.2

5, 2

.0]

Any

info

rmal

vsn

one

Æ.8

6[.4

1, 1

.77]

.77

[.36,

1.6

3].8

2[.3

3, 2

.08]

.77

[.30,

1.8

4].6

4[.2

4, 1

.75]

.59

[.22,

1.5

8]

Citi

zens

hip

Non

evs

som

eÆ

1.06

[.79,

1.4

2].9

4[.6

8, 1

.30]

.82

[.56,

1.2

1].7

6[.5

0, 1

.15]

1.23

[.80,

1.9

0]1.

22[.7

9, 1

.90]

Co

GN

ITIv

e So

CIA

l C

ApI

TAl

Med

ium

vsl

owÆ

.56

[.29,

1.1

0].6

4[.3

1, 1

.31]

.54

[.24,

1.2

2].5

5[.2

3, 1

.28]

.18*

**[.0

7, .4

7].1

8***

[.07,

.47]

Hig

hvs

low

Æ.4

5***

[.25,

.78]

.57

[.32,

1.0

5].4

4*[.2

3, .8

6].5

0[.2

4, 1

.00]

.30*

**[.1

5, .5

9].3

2***

[.16,

.64]


21

Tabl

e 9

cont

inue

d…

eigh

t-ye

ar-o

lds’

heal

th o

utco

mes

Soci

al ca

pita

lIl

lnes

s in

last

two

wee

ks(y

es =

1)

Men

tal h

ealth

scor

es >

=17

(yes

= 1

)

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d a

odds

ratio

[95%

CI]

Cru

deod

ds ra

tio[9

5%C

I]

Adj

uste

d a

odds

ratio

[95%

CI]

STR

UC

TU

RA

l So

CIA

l C

ApI

TAl

Gro

up m

embe

rshi

p

Any

vsn

one

Æ.9

7[.7

3, 1

.30]

1.08

[.80,

1.4

8]1.

19[.8

5, 1

.66]

1.29

[.90,

1.8

5]

Any

form

alv

snon

eÆ

1.0

[.73,

1.3

3]1.

14[.8

2, 1

.58]

1.19

[.84,

1.6

8]1.

37[.9

5, 2

.0]

Any

info

rmal

vsn

one

Æ.8

6[.7

3, 1

.37]

.80

[.45,

1.4

0]1.

56[.8

8, 2

.79]

1.28

[.68,

2.1

0]

Soci

al su

ppor

t rec

eive

d

Any

vsn

one

Æ1.

75[.7

5, 4

.08]

1.78

[.75,

4.2

5].6

1[.2

8, 1

.35]

.40*

[.18,

.89]

Any

form

alv

snon

eÆ

1.53

[.64,

3.6

3]1.

62[.6

7, 3

.95]

.55

[.24,

1.2

4].3

4*[.1

5, .8

0]

Any

info

rmal

vsn

one

Æ1.

75[.7

5, 4

.07]

1.77

[.74,

4.1

2].6

1[.2

8, 1

.35]

.40*

[.18,

.89]

Citi

zens

hip

Som

evs

non

eÆ

1.23

[.92,

1.6

4]1.

36[1

.0, 1

.86]

.69*

[.49,

.96]

.64*

[.44,

.92]

Co

GN

ITIv

e So

CIA

l C

ApI

TA

l

Med

ium

vsl

owÆ

.83

[.43,

1.6

0].9

9[.4

9, 1

.98]

.86

[.43,

1.7

3]1.

15[.5

4, 2

.42]

Hig

hvs

low

Æ.4

8**

[.27,

.83]

.58

[.32,

1.0

6].3

9*[.2

2, .7

2].4

5*[.2

4, .8

5]

a A

djus

ted

for:

(1)

Mot

hers

: ag

e, e

duca

tion

leve

l, le

ngth

of

livin

g in

com

mun

e, e

thni

c gr

oup,

occ

upat

ion,

men

tal h

ealth

sta

tus;

(2)

Hou

seho

ld:

wea

lth in

dex,

sho

ck e

vent

s in

last

thre

e ye

ars;

(3)

C

omm

unity

: ru

ral/u

rban

; (4)

Chi

ld: s

ex, a

ge

* p<

0.05

; **

p<0.

01; *

** p

<0.

001

(com

pare

d w

ith b

asel

ine

grou

p)


22

Tabl

e 10

: Log

isti

c re

gres

sion

est

imat

es o

f pre

dict

ors

of m

ater

nal s

truc

tura

l soc

ial c

apit

al (

soci

al s

uppo

rt)

of m

othe

rs o

f one

-yea

r-ol

ds

Fact

ors

Stru

ctur

al so

cial

capi

tal:

Rec

eive

d su

ppor

t fro

m

Any v

s Non

eaAn

y for

mal

vs N

oneb

Any i

nfor

mal

vs N

oneb

Cru

de1

odds

ratio

[95%

CI]

Adj

uste

d 2od

ds ra

tio[9

5%C

I]

Cru

de1

odds

ratio

[95%

CI]

Adj

uste

d 2od

ds ra

tio[9

5%C

I]

Cru

de1

odds

ratio

[95%

CI]

Adj

uste

d 2od

ds ra

tio[9

5%C

I]

Mo

TH

eR

Age

(in

year

s)1.

00[.9

5, 1

.04]

.98

[.94,

1.0

3]1.

02[1

.00,

1.0

8]1.

00[.9

5, 1

.05]

1.00

[.95,

1.0

4].9

8[.9

4, 1

.03]

leng

th o

f liv

ing

in co

mm

une (

in ye

ars)

1.00

[.97,

1.0

1]1.

00[.9

7, 1

.02]

1.00

[.97,

1.0

1]1.

00[.9

7, 1

.02]

1.00

[.97,

1.0

1]1.

00[.9

7, 1

.02]

ethn

icit

y (0

= ot

hers

; 1 =

kin

h)2.

01*

[1.1

3, 3

.58]

1.15

[.60,

2.2

3]2.

10*

[1.1

4, 3

.88]

.93

[.46,

1.9

0]2.

02*

[1.1

4, 3

.61]

1.17

[.60,

2.5

6]

educ

atio

n le

vel

Prim

ary

vsn

osc

hool

ing

Æ1.

80[.9

7, 3

.29]

1.42

[.71,

2.8

5]1.

50[.8

0, 2

.82]

1.14

[.55,

2.3

7]1.

80[1

.00,

3.2

9]1.

42[.7

1, 2

.84]

Seco

ndar

yvs

no

scho

olin

gÆ

1.77

[.96,

3.2

7]1.

03[.4

5, 2

.39]

2.42

**[1

.28,

4.5

7]1.

24[.5

2, 2

.95]

1.77

[.96,

3.2

5]1.

03[.4

4, 2

.38]

occ

upat

ion

(0 =

oth

er; 1

= ag

ricu

lture

).5

6*[.3

4, .9

4].7

5[.4

1, 1

.39]

.49*

*[.2

9, .8

3].7

3[.3

9, 1

.37]

.56*

[.34,

94]

.75

[.41,

1.3

8]

Men

tal h

ealth

stat

us (0

= n

o; 1

= ye

s).5

0*[.2

9, .8

6].4

1**

[.23,

.72]

.57

[.33,

1.0

2].4

7**

[.26,

.85]

.50*

[.29,

.85]

.40*

*[.2

3, .7

2]

Ho

USe

Ho

lD

Wea

lth in

dex

<0.5

(0 =

no;

1 =

yes)

1.61

[.94,

2.7

5].6

7[.3

1, 1

.46]

2.11

**[1

.21,

3.6

7]1.

75[.3

4, 1

.67]

1.61

[.94,

2.7

6].6

7[.3

1, 1

.46]

Shoc

k ev

ents

hap

pene

d in

last

thre

e yea

rs

1ev

entv

snon

eÆ

1.32

[.71,

2.4

4]1.

53[.7

8, 3

.00]

1.30

[.69,

2.4

5]1.

50[.7

5, 3

.00]

1.33

[.72,

2.4

5]1.

55[.7

9, 3

.04]

>=2

even

tsvs

non

eÆ

1.04

[.50,

2.1

4]1.

55[.7

3, 3

.30]

1.32

[.63,

2.7

7]2.

06[.9

5, 4

.50]

1.

03[.5

0, 2

.13]

1.55

[.73,

3.3

0]


23

Fact

ors

Stru

ctur

al so

cial

capi

tal:

Rec

eive

d su

ppor

t fro

m

Any v

s Non

eaAn

y for

mal

vs N

oneb

Any i

nfor

mal

vs N

oneb

Cru

de1

odds

ratio

[95%

CI]

Adj

uste

d 2od

ds ra

tio[9

5%C

I]

Cru

de1

odds

ratio

[95%

CI]

Adj

uste

d 2od

ds ra

tio[9

5%C

I]

Cru

de1

odds

ratio

[95%

CI]

Adj

uste

d 2od

ds ra

tio[9

5%C

I]

livi

ng ar

ea (0

= ru

ral;

1 =

urba

n)1.

36*

[1.0

7, 1

.74]

1.35

*[1

.04,

1.7

4]1.

47**

[1.1

5, 1

.87]

1.39

**[1

.07,

1.8

1]1.

36*

[1.0

7, 1

.74]

1.35

*[1

.04,

1.7

5]

CH

IlD

Chi

ld’s

sex

(0 =

gir

l; 1

= bo

y)1.

33[.8

0, 2

.21]

1.45

[.86,

2.4

4]1.

30[.7

7, 2

.21]

1.46

[.85,

2.5

1]1.

34[.8

1, 2

.23]

1.46

[.87,

2.4

6]

Age

(in

mon

ths)

.06

[.88,

1.0

5].9

5[.8

7, 1

.04]

.96

[.88,

1.0

5].9

5[.8

6, 1

.04]

.96

[.88,

1,0

5].9

5[.8

7, 1

.04]

HU

SBA

ND

part

ner’s

educ

atio

n

Prim

ary

vsn

osc

hool

ing

Æ1.

39[.7

6, 2

.53]

1.21

[.62,

2.3

4]1.

47[.7

7, 2

.79]

1.36

[.67,

2.7

4]1.

39[.7

6, 2

.54]

1.21

[.62,

2.3

4]

Seco

ndar

yvs

no

scho

olin

gÆ

2.88

***

[1.5

3, 5

.43]

2.59

*[1

.23,

5.4

4]4.

22**

*[2

.17,

8.2

0]3.

34*

[1.5

3, 7

.27]

2.87

***

[1.5

2, 5

.41]

2.57

*[1

.22,

5.4

0]

a lo

gist

ic m

odel

(ba

se c

ateg

ory

= n

one

or n

ot r

ecei

ving

any

sup

port

from

indi

vidu

al a

s w

ell a

s gr

oup

netw

ork)

, b M

ultin

omia

l log

istic

mod

el (

base

cat

egor

y =

not

rec

eivi

ng a

ny s

uppo

rt)

1 n

= 19

53; 2

n =

192

6 (2

7 m

issi

ng o

bser

vatio

ns)

* p<

0.05

; **

p<0.

01; *

** p

<0.

001

(com

pare

d w

ith b

asel

ine

grou

p)


24

Table 11: Logistic regression estimates of predictors of maternal cognitive social capital of mothers of one-year-olds

Factors

Cognitive social capitalb

Medium vs low High vs low

Crude1odds ratio[95%CI]

Adjusted 2odds ratio[95%CI]

Crude1odds ratio[95%CI]

Adjusted 2odds ratio[95%CI]

MoTHeR

Age (in years) 1.03[1.00, 1.07]

1.03[1.00, 1.08]

1.02[.99, 1.05]

1.03[.99, 1.06]

length of living in commune (in years) 1.02[1.00, 1.03]

1.00[.99, 1.03]

1.02*[1.00, 1.03]

1.00[.99, 1.02]

ethnicity (0 = others; 1 = kinh) .83[.46, 1.48]

.69[.34, 1.40]

.76[.47, 1.23]

.69[.38, 1.23]

education level

PrimaryvsnoschoolingÆ.97

[.59, 1.60].95

[.53, 1.71].95

[.64, 1.41].89

[.55, 1.42]

SecondaryvsnoschoolingÆ1.41

[.84, 2.35]1.21

[.61, 2.40]1.08

[.71, 1.64].89

[.51, 1.56]

occupation (0 = other; 1 = agriculture) 1.15[.77, 1.70]

1.04[.66, 1.64]

.86[.62, 1.17]

.67*[.46, .99]

Mental health status (0 = no; 1 = yes) .62[.41, .96]

.66[.42, 1.04]

.34***[.24, .48]

.38***[.27, .55]

HoUSeHolD

Wealth index <0.5 (0 = no; 1 = yes) 1.09[.72, 1.64]

1.22[.69, 2.13]

1.17[.84, 1.64]

1.43[.89, 2.29]

Shock events happened in last three years

1eventvsnoneÆ.79

[.50, 1.24].83

[.51, 1.35].74

[.51, 1.08].85

[.57, 1.27]

>=2eventsvsnoneÆ.85

[.49, 1.46].94

[.53, 1.67].73

[.47, 1.13].91

[.57, 1.45]

living area ( 0 = rural; 1= urban) .92[.82, 1.03]

.87[.75, 1.01]

0.90*[.82, .99]

0.81**[.72 , .93]

CHIlD

Child’s sex (0 = girl; 1=boy) 1.35[.90, 2.00]

1.35[.90, 2.04]

1.25[.90, 1.73]

1.29[.92, 1.80]

Age (in months) 1.00[.94, 1.07]

1.00[.94, 1.08]

.99[.94, 1.04]

.99[.94, 1.05]

HUSBAND

partner’s education

PrimaryvsnoschoolingÆ1.19

[.69, 2.07]1.34

[.75, 2.39].98

[.64, 1.51]1.09

[.68, 1.75]

SecondaryvsnoschoolingÆ1.45

[.87, 2.43]1.43

[.76, 2.70]1.12

[.75, 1.67]1.18

[.70, 1.99]

a Multinomial logistic model (base category = low level) 1 n = 1953; 2 n = 1926 (27 missing observations) * p<0.05; ** p<0.01; *** p<0.001 (compared with baseline group)


25

4. discussion and conclusionsAswithallcross-sectionalstudies,causalitycannotbeattributedtotheassociationsfoundinthispaper.Inaddition,theover-samplingofpoorrespondentsfortheYoungLivesstudy,andthemeasurementofonlymaternalsocialcapital,meansthattheseresultscannotbegeneralisedtothepopulationofVietnamasawhole.

Thisstudyonlymeasuresthesocialcapitalofachild’smother.However,childrenarelikelytobeinfluencedbythesocialcapitalofothermembersofthehousehold,andindeedacompositemeasureof‘householdsocialcapital’mayshowstrongerassociationswithchildhealth.Theplaceofwomeninsocietymayaffectaccesstosocialcapital,forexamplebybeingamemberoffewerformalorganisationsorsocialgroupsthanmen.Inaddition,somemeasurementofthechild’ssocialcapital,particularlyoftheeight-year-olds,mayhelptoexplainsomeaspectsofchildhealth.

Fourmainconclusionscanbedrawnfromthisstudy.Firstly,maternalsocialsupportandhighcognitivesocialcapitalshowthemostconsistentassociationswithchildhealth.Thefactthatthesevariablesrelatetoamorenurturingenvironmentinwhichthechilddevelopsmayexplaintheimportanceofthesevariablesforchildhealth.

Secondly,associationsbetweenmaternalsocialcapitalandchildhealtharemuchstrongeramongone-year-oldchildrenthanamongeight-year-olds,withtheexceptionofchildmentalhealthandlife-threateningillness.Wehypothesisethatthisisbecauseinfantshavemuchgreatercontactwiththeirmother,andarethereforeinfluencedbythesocialworldinwhichshemovestoagreaterextent.Perhapsbytheageofeight,childrenhavebeguntodeveloptheirownsocialcapital,thoughtheiremotionaldevelopmentmaystillbeinfluencedbythesocialcapitaloftheirmother.

Thirdly,themostconsistentassociationsareseenwiththementalhealthofeight-year-olds,withthreeofthefourindicatorsofsocialcapitalbeingsignificantlyassociatedwithmentalhealthafterothervariableshavebeencontrolledfor.Otherstudieshaveshownsocialcapitaltobemoreconsistentlyassociatedwithmentalratherthanphysicalhealth(DeSilvaetal,2005);thequalityofthesocialenvironmentisperhapsexpressedmorereadilythroughemotionalthanthroughphysicalhealth.

Lastly,thereissomeevidencetosuggestthatactivemembershipofformalorganisationsinVietnammaybedamagingtothehealthofeight-year-oldsbutnotone-year-olds.Activeparticipationinformalgroupswassignificantlyassociatedwithanincreaseinstuntingamongeight-year-olds.Withthiscross-sectionaldataitisnotpossibletodeterminewhetherthecostsimposedonmothersparticipatinginformalgroupsleadtochildstunting,orwhethermotherswithastuntedchildaremorewillingtoparticipateinformalgroupsinordertoreceivesocialsupportfromthosegroups,thoughthiscanbeexploredinfutureroundsoftheYoungLivesproject.However,onepreviousstudydidfindanegativeimpactofmembershipofgroupsonmentalhealth,indicatingthepotentiallydamagingsidetogroupmembership(MitchellandLaGory,2002).Groupmembershipcandamagementalhealthbybeinganextraburdenontopoftheproductiveandreproductiverolesofwomenand/oritcanproducestigmaandpeerpressure,whichincreasestressorsuponawoman.Thishypothesiscouldbeusefullyexploredbyqualitativeresearchwhichtakescasesofmotherswithhighlevelsofparticipationwhohave


2�

malnourishedchildren.Themothers’perceptionsofparticipationasaburden/choreversusapositiveopportunitycouldbeexamined.

Structuralsocialcapitalisremarkablylowinacontextwheremassorganisationsstillexist.TheonlyotherquantitativestudyofsocialcapitalinVietnamfoundhigherlevelsofstructuralsocialcapitalbutitincludedmeninitssample.However,whenanalysingactivemembershipofgroups(aswasdoneinthisstudy)ageneral‘socialdisconnectedness’wasfound(DaltonandOng,2001).Thoughwomen’sunionsremainanimportantformalnetworkinVietnamwithapproximately20percentofmothersactivemembers,othertypesofinformalandformalgroupshaveamuchsmallerinfluence,inparticularinformalorganisationssuchasreligiousandsportsgroups.Theonlyassociationfoundinthisstudybetweenactivemembershipofgroupsandchildhealthwaswithnutritionalstatusofeight-year-oldsmeasuredbyheight-for-agez-scores,butthisstatisticalcorrelationhaslesspracticalimplications.

Apositiveassociationbetweensocialsupportfrominformalnetworksandthenutritionstatusofone-year-oldchildren(weight-for-age)wasfound.Thesenetworkscomprisefamily,friendsandneighboursandareastrongfeatureofVietnamesesociety,with96percentofmothersreceivingsupportfromthissource.For65percentofmothers,thisconstitutestheironlysourceofsupport.ConfuciantraditionsoftenpromotetrustinarelativelynarrowrealmoffamilyandtheVietnamesefamilyhasbeencharacterisedas‘residentiallynuclearbutfunctionallyextended’(Jones,1995p.189),withextensivesupportofferedbythefamily(Pham,1999).Daltonetal(2002)hypothesisethatdevelopmentinVietnamwillnotdiminishtraditionalfamilynetworksbutaddworkandfriendshipnetworkstothem:‘furtherdevelopmentinVietnamisnotsolikelytoexchangeonesetofsocialnetworksforanother,buttoexpandthenumberandactivitylevelsofthenetworksthatconnectindividualstosociety,andwhichhelpformtheirsocialandpoliticalidentities’(p.4).OnlylongitudinalresearchwillbeabletoaddressthishypothesisandYoungLivesiswellplacedtodoso.

Somepreviousstudieshaveshownlevelsofadults’generalisedtrusttobegenerallylowindevelopingcountries(ProductivityCommission,2003).Indeed,preliminaryresultsfromPeru,oneoftheothercountriestakingpartintheYoungLivesstudy,showlevelsofmaternalgeneralisedtrustinthecommunitytobeonly30percent(www.younglives.org.uk).Incontrast,Vietnamhasextremelyhighlevelsoftrust,with84percentofmothersbelievingthatpeopleingeneralcanbetrusted.Thisisechoedinthehighlevelsoftheotherindicatorsofcognitivesocialcapital:senseofcommunityharmony,senseofbelonging,andperceivedfairness.DaltonandOng(inpress),usingsimilarquestions,alsofoundhighlevelsoftrustinVietnam–higherthaninChinaandtheworldaverage.Highmaternalcognitivesocialcapitalisassociatedwithbetterchildhealthintermsofnutritionstatusandphysicalhealthamongone-year-olds,andinphysicalandmentalhealthamongeight-year-olds.

AsthisisonlythesecondstudytoconsidersocialcapitalinVietnam,andthefirsttostudymaternalsocialcapitalandchildwell-being,anyattempttoidentifypolicyimplicationsmustbeconsideredexploratory.Inaddition,thisstudywasnotabletoassesstherelativeimportanceofmaternalsocialcapitalasariskfactorforchildhealthasitdoesnotincludeallpotentialriskfactorsforchildhealth.Forexample,therelativeimportanceofmaternalsocialcapitalasariskfactorforchildmalnutritioncannotbeanalysedbecausewewereunabletoincludeallpossibleriskfactorsformalnutritioninthemodels(forexample,YoungLivesdoesnotmeasurefoodavailability).Asthisstudyhasestablishedthataspectsofmaternalsocialcapitalareassociatedwithsomechildhealthoutcomes,additionalstudies


27

areneededtoassessitsimportancerelativetootherknownriskfactorsforchildhealth(forexamplepoverty)andthereforeitspolicyimportance.However,raisingstructuralsocialcapital(butensuringthisdoesnotcreatedamagingburdenstowomen)andmaintainingthehighlevelsofcognitivesocialcapitalisobviouslyageneralpolicypriority.StrengtheningsocialcapitalinVietnamcanbeenvisagedintwo(mutuallysupporting)scenarios.Firstly,inaformalisingscenarioitisrecognisedthattraditionalsocialdynamicsarechanging(throughurbanisation,reductionofextendedfamilyandageingpopulation)andthatformalstructures,eg,massandnon-governmentalorganisationsandcommunity-basedorganisations(CBOs),willneedtostepintofacilitatesocialnetworksandsupport.Thoughtheirroleshavebeenincreasinglyrecognisedbythegovernmentandinthepublicdiscourse,thedevelopmentoftheseorganisationsis,however,constrainedbyanunsupportivelegalenvironment.Agovernmentdecreein2003(88/2003NC-CP)withregardstotheorganisationandmanagementofpeople’sassociationsinVietnamgivesapositivesignal.However,itfailstogiveclearoperationalguidelines,whichshouldbeapriorityforpolicydebateandformulation.Inaddition,givenaclearroleofengaginginsocialmobilisationworkandsecuringbenefitsfortheirownmembers,massorganisationscouldperformtheirrolebetterbytransformingthemselvesinordertoadapttosocialchanges.Thiscanbedone,andisbeingdoneinsomemassorganisations,bygoingbeyondmobilisingsupportforthecentralCommunistPartytomeetingtheneedsof,andadvocatingonbehalfof,theirmembership.Forexample,theyneedtodiversifytheiractivitiesinsuchawaythatcouldinterest,andmeetthevariousneedsof,theirmembers.Anexampleofthewomen’sunion’sdiversificationofactivitiesisthatthemeetingofwomen’suniongroupsshouldnotfocusonlyondiscussingtheirownproceduralactivitiesbutalsointegrateotheractivities(suchaslifeskillstraining,familyplanning,HIV/AIDS,businessplanning,etc)andcapacity-buildingfortheirlocalstaffinordertohandletheirnewrole.

Secondly,aninformalisingscenarioenvisagesnewindividualconnectionsandsupportarisingfromthenewsocialstructures.Strategiestorespondtothisscenariorequireactionsaimedatsupportinginformalstructuresratherthanformalinstitutions.InVietnam,thereisasupportivepolicyenvironmentforindividualstostrengthentheirinformalsocialcapitalandgeneralcitizenship.Theintroductionof‘Decree29’3anditsrecentrevisionthatwasbuiltonaprincipleof‘peopleknow,peoplediscuss,peoplemanage,andpeoplesupervise’indicatesgovernmentsupportinindividualcapacity-buildingandfacilitatinglocalpeople’sparticipationindecision-making.Accordingtothedecree,peopleareencouragedtodiscussandmakedecisionsonissuesthatconcernoraffectthem.Inaddition,theyareencouragedtoprovidefeedbackondecisionsmadebyrelatedstakeholdersandhigherauthorities.TheroleofCBOs/people’sorganisationswouldfitwellinthiscontextastheycouldfacilitatelocalpeopletospeakforthemselves,orlinkpeoplewithhigher-levelauthorities.Oneofthebiggestchallengesinimplementingthedecreeisthelackofpracticalguidanceandthedisincentiveoflossofpowerforlocalcivilservants(WorldBankandADB,2002).

ThisfirstquantitativestudyofmaternalsocialcapitalinVietnamhasshownittobeassociatedwithsomeaspectsofchildhealth.PolicydebateonstrengtheningsocialrelationsinadynamiccontextneedstobeinformedbythisdataandbyfutureroundsoftheYoungLivesstudythatwillproducedatacapableofexploringcausalrelationshipsaswellasthechangingsocialsituation.

3 Thisdecreeaimstostrengthentheparticipationofgrassrootspeopleinlocalplanninganddecision-making.


28

referencesArrow,K(2000)‘Observationsonsocialcapital’,inPDasguptaandISerageldin(eds)Social Capital: A multifaceted perspective,Washington,DC,WorldBank

Bourdieu,P(1986)‘Theformsofsocialcapital’,inJRichardson(ed)The Handbook of Theory and Researchfor the Sociology of Education, NewYork,GreenwoodPress

Coleman,JS(1990)Foundations of Social Theory,Cambridge,Mass.,BelknapPressofHarvardUniversityPress

Conway,TandCTurk(2002)Reducing Vulnerability and Providing Social Protection,Hanoi,PovertyTaskForce

Dalton,RandTOng(2001)The Vietnamese Public in Transition: The World Values Survey 2001,IrvineUSA,CenterfortheStudyofDemocracy

Dalton,R,PMinhHac,PThanhNghiandTOng(2002)‘SocialrelationsandsocialcapitalinVietnam:The2001WorldValuesSurvey’,Comparative Sociology,1,369-386

Dalton,RandTOng(inpress)‘CivilsocietyandsocialcapitalinVietnam’, inModernization and Social Change in Vietnam,Munich,MunichInstituteforSocialScience.www.democ.uci.edu/democ/papers/vietnam04.pdf

Dasgupta,PandISerageldin(eds)(2000)Social Capital: A multifaceted perspective,Washington,DC,WorldBank

DeSilva,M,KMcKenzie,SHuttlyandTHarpham(2005)‘Mentalillnessandsocialcapital:Asystematicreview’,Journal of Epidemiology and Community Health, 59,619-627

GoV(2003)Vietnam Growth and Reduction of Poverty: Annual progress report 2002-03,Hanoi,SteeringCommitteeoftheComprehensivePovertyReductionandGrowthStrategy(CPRGS),Inter-MinisterialWorkingGroup

Grootaert,CandTvanBastelaer(2002)Understanding and Measuring Social Capital: Amultidisciplinary tool for practitioners,Washington,DC,WorldBank

Harpham,T,EGrantandEThomas(2002)‘Measuringsocialcapitalwithinhealthsurveys:Somekeyissues’,Health Policy and Planning,17(1),106-111

JointDonorReport(2003)Vietnam Development Report 2004,Hanoi,JointDonorReporttotheVietnamConsultativeGroupMeeting

Jones,GW(1995)‘PopulationandfamilyinSoutheastAsia’,Journal of Southeast AsianStudies, 26(1),184-195

Mitchell,CandMLaGory(2002)‘Socialcapitalandmentaldistressinanimpoverishedcommunity’,City and Community,1(2),199-222

Pedersen,K(2001)Civil Society in the Context of Development Aid: The case of Vietnam,Copenhagen,CopenhagenBusinessSchoolWorkingPaper


2�

Pham,VB(1999)The Vietnamese Family in Change: The case of the Red River Delta,Surrey,UK,CurzonPress

Portes,A(1998)‘Socialcapital:Itsoriginsandapplicationsinmodernsociology’,Annual Review of Sociology,24,1-24

ProductivityCommission(2003)Social Capital: Reviewing the concept and its policy implications, Canberra,CommonwealthofAustralia

Putnam,R(1993)‘Theprosperouscommunity:Socialcapitalandpubliclife’,The American Prospect,4(13),27-40.www.prospect.org/print/V4/13/putnam-r.html

Quibria,M(2003)The Puzzle of Social Capital: A critical review,ERD(EconomicsandResearchDepartment)WorkingPaperNumber40,Manila,AsianDevelopmentBank

SavetheChildrenUK(2003)HCMC Participatory Poverty Assessment,London,SavetheChildrenUK/PovertyTaskForce

Shanks,EandCTurk(2002)Policy Recommendations from the Poor,Hanoi,PovertyTaskForce

StataCorporation(2003)Stata Statistical Software: Release 8.0,CollegeStation,TX,Stata

Tuan,T,THarpham,NTHuong,MDeSilva,VTTHuong,TTLong,NTVHaandDDewitt(2005)Measuring Social Capital and Mental Health in Vietnam: A validity study,YoungLivesWorkingPaper12

UNCT (2003) MDGs Closing the Millennium Gaps: Progress Report 2003,Hanoi,UNCT(UnitedNationsCountryTeam)

VanKemenade,S(2002)Social Capital as a Health Determinant: How is it defined?GovernmentServicesCanada,HealthPolicyResearchWorkingPaperSeries

Wischermann,JandNQVinh(2003)'TherelationshipbetweencivicorganizationsandgovernmentalorganizationsinVietnam',inGetting Organized in Vietnam,Singapore,InstituteofSouthEastAsianStudies

WorldBank(2003)Vietnam Development Report: Poverty,Hanoi,WorldBank

WorldBankandADB(2002)Vietnam, Delivering on its Promise,Hanoi,WorldBankandAsianDevelopmentBank


30

The Young Lives Partners Centre for Economic and Social Studies (CESS), India

Department of Economics, University of Addis Ababa, Ethiopia

Ethiopian Development Research Institute, Addis Ababa, Ethiopia

General Statistical Office, Government of Vietnam

Grupo de Análisis Para El Desarrollo (GRADE), Peru

Institute of Development Studies, University of Sussex, UK

Instituto de Investigación Nutricional (IIN), Peru

London School of Hygiene and Tropical Medicine, UK

Medical Research Council of South Africa

Queen Elizabeth House, University of Oxford

RAU University, Johannesburg, South Africa

Research and Training Centre for Community Development, Vietnam

Save the Children UK

South Bank University, UK

Statistical Services Centre, University of Reading, UK

W O R K I N G P A P E R N O . 3 0

Maternal Social Capital and Child Health in VietnamTran TuanTrudy HarphamMary J De SilvaNguyen Thu HuongBill TodPham Thi LanTran Duc ThachSavitri Abeyasekera

In Vietnam there is growing concern about the potential social impact of rapid economic changes. The extent and type of social connectedness, or social capital, may be changing. Studies from other developing countries have demonstrated that social capital is often independently associated with various indicators of well-being, including some aspects of human capital (health and education status). In Vietnam there has only been one previous quantitative study of social capital and this did not consider associations with well-being. The Young Lives project in Vietnam allows the examination of the relationship between maternal social capital and child well-being. With a sample of 1,953 mothers of one-year-olds and 954 mothers of eight-year-olds across five provinces, this study examines whether maternal social capital is associated with child health.

This study found low levels of structural social capital (associational life) and citizenship and high levels of cognitive social capital (trust, etc) and support. While poorer women have even lower levels of structural social capital than their better-off peers, patterns across rural and urban areas are similar.

Maternal social support and high cognitive social capital show the most consistent associations with child health, while associations between maternal social capital and child health are stronger among one-year-old children than among eight-year-olds, with the exception of child mental health and risk of life-threatening illness. The most consistent associations were seen with the mental health of eight-year-olds, with all four indicators of social capital being significantly associated with mental health after other variables have been controlled for. Lastly, there was some evidence to suggest that active membership of formal organisations in Vietnam may be damaging to the health of eight-year-olds but not one-year-olds, with active participation in formal groups being significantly associated with an increase in stunting among eight-year-olds.

The main components of social capital associated with child health are social support and cognitive social capital. High levels of maternal social support are positively correlated with having an educated partner, living in an urban area and better maternal mental health. High cognitive social capital is positively correlated with living in a rural area, having an occupation other than agriculture and better maternal mental health.

There is a need to focus on maintaining the high levels of cognitive social capital and raising the current low level of maternal structural social capital in Vietnam.

Key words: social capital, child health, Vietnam, quantitative

Published by

Young Lives Save the Children UK 1 St John's Lane London EC1M 4AR

Tel: 44 (0) 20 7012 6796 Fax: 44 (0) 20 7012 6963 Web: www.younglives.org.uk

ISBN 1-904427-31-6 First Published: 2005

All rights reserved. This publication is copyright, but may be reproduced by any method without fee or prior permission for teaching purposes, though not for resale, providing the usual acknowledgement of source is recognised in terms of citation. For copying in other circumstances, prior written permission must be obtained from the publisher and a fee may be payable.

Designed and typeset by Copyprint UK Limited

Young Lives is an international longitudinal study of childhood poverty, taking place in Ethiopia, India, Peru and Vietnam, and funded by DFID. The project aims to improve our understanding of the causes and consequences of childhood poverty in the developing world by following the lives of a group of 8,000 children and their families over a 15-year period. Through the involvement of academic, government and NGO partners in the aforementioned countries, South Africa and the UK, the Young Lives project will highlight ways in which policy can be improved to more effectively tackle child poverty.

maternal social capital and child health in vietnam

Documents