looking upstream: enhancers of child nutritional status in …llanes,[email protected],...

Submitted 2 December 2015Accepted 6 February 2016Published 1 March 2016

Corresponding authorJose Manuel Rodriguez-Llanes jmrllanesgmailcomjoserodriguezuclouvainbe

Academic editorJara Peacuterez-Jimeacutenez

Additional Information andDeclarations can be found onpage 17

DOI 107717peerj1741

Copyright2016 Rodriguez-Llanes et al

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

Looking upstream enhancers of childnutritional status in post-flood ruralsettingsJose Manuel Rodriguez-Llanes1 Shishir Ranjan-Dash23 Alok Mukhopadhyay4

and Debarati Guha-Sapir1

1Centre for Research on the Epidemiology of Disasters Institute of Health and SocietyUniversiteacute Catholique de Louvain Brussels Belgium

2Department of Management Siksha lsquoOrsquo Anusandhan University Bhubaneswar India3Tata Trusts Mumbai India4Voluntary Health Association of India New Delhi India

ABSTRACTBackground Child undernutrition and flooding are highly prevalent public healthissues in many developing countries yet we have little understanding of preventivestrategies for effective coping in these circumstances Education has been recentlyhighlighted as key to reduce the societal impacts of extreme weather events underclimate change but there is a lack of studies assessing to what extent parental educationmay prevent post-flood child undernutritionMethods andMaterials One year after large floods in 2008 we conducted a two-stage cluster population-based survey of 6ndash59 months children inhabiting floodedand non-flooded communities of Jagatsinghpur district Odisha (India) and collectedanthropometric measurements on children along with child parental and householdlevel variables through face-to-face interviews Using multivariate logistic regressionmodels we examined separately the effect of maternal and paternal education andother risk factors (mainly income socio-demographic and child andmother variables)on stunting and wasting in children from households inhabiting recurrently floodedcommunities (2006 and 2008 n= 299) As a comparison separate analyses on childrenin non-flooded communities were carried out (n= 385) All analyses were adjusted byincome as additional robustness checkResults Overall fathers with at least completed middle education (up to 14 years ofage and compulsory in India) had an advantage in protecting their children from childwasting and stunting For child stunting the clearest result was a 100ndash200 lowerprevalence associated with at least paternal secondary schooling (compared to noschooling) in flooded-areas Again only in flooded communities an increase in percapita annual household income of 1000 rupees was associated to a 47ndash49 lowerprevalence of child stunting For child wasting in flooded areas delayed motherhoodwas associated to better nutritional outcomes (34 lower prevalence per year) Inflooded communities households dedicated to activities other than agriculture a 50ndash51 lower prevalence of child wasting was estimated suggesting farmers and fishermenas the most vulnerable livelihoods under flooding In flooded areas lower rank casteswere at higher odds of both child wasting and stuntingConclusions In the short-term protracted nutritional response in the aftermath offloods should be urgently implemented and target agricultural livelihoods and low-rank castes Education promotion and schooling up to 14 years should have positive

How to cite this article Rodriguez-Llanes et al (2016) Looking upstream enhancers of child nutritional status in post-flood rural set-tings PeerJ 4e1741 DOI 107717peerj1741

impacts on improving children nutritional health in the long run especially underflooding Policies effectively helping sustainable livelihood economic development anddelayed motherhood are also recommended

Subjects Epidemiology Global Health Nutrition Pediatrics Public HealthKeywords Flood Climate change adaptation Wasting Stunting Father education Mothereducation Parental education Income Maternal autonomy Subsistence farming

INTRODUCTIONAmong all the disaster risks associated to a warming climate flooding has become themost frequent and since the nineties it has been affecting grosso modo 100 million peoplea year This is more than any other disaster type worldwide climate-related or not(EM-DAT 2015) Yet relative to their importance the health consequences of floodinghave been rarely investigated (Ahern et al 2005 Alderman Turner amp Tong 2012) Even ifat the time of writing an equivalent 378 (ie 28 billion people) of the current worldpopulation were affected by floods in the past 25 years little serious actions are being taken

Education has been recently stressed as one key to reduce the societal impacts ofclimate change (Striessnig Lutz amp Patt 2013) The study of educational attainment as apromotor of disaster resilience recently received substantive attention in a collection of11 papers (Muttarak amp Lutz 2014) The results of this special issue are a convincing stepforward These studies were undertaken in different countries and regions and targetedrelevant outcomes in different phases of the disaster cycle in populations exposed to diverseclimate-related hazards Overall they found that higher-educated groups avoided high-riskareas to settle were better prepared reacted more efficiently to early warnings and hadlessen impacts on health and social variables which indicated according to the authorsbetter coping and recovery in these groups (Muttarak amp Lutz 2014) For the particular caseof flooding the evidence on the role of education as a promotor of positive health outcomesis scant and contentious (Lowe Ebi amp Forsberg 2013) But what are the plausible pathwayslinking more education to health improvements Indeed formal education is crucialfor acquisition and processing of information (eg literacy) improvement of cognitiveabilities decision making and long-term planning and generally it leads to securing skilledjobs and ultimately higher income and better health (briefly reviewed in Striessnig Lutz ampPatt 2013 Lindeboom Llena-Nozal amp Van der Klauuw 2009)

The transgenerational effects of education on health have received substantial attentionAlthough the effect of maternal education on child stunting and general child healthhas been studied (Milman et al 2005 Lindeboom Llena-Nozal amp Van der Klauuw2009) the effect of paternal education on child health has been more rarely examined(Moestue amp Huttly 2008 Semba et al 2008) Assessing the extent to which the educationof the father affects a childrsquos health is an important step to understand the relativecontribution of fathers to the family wellbeing Considering the fatherrsquos education is alsoa premise to further study the synergies between father and mother education levels in thepromotion of childrsquos health (Semba et al 2008)

Rodriguez-Llanes et al (2016) PeerJ DOI 107717peerj1741 221

Adequate child nutrition is a key indicator of wellbeing and development of particularimportance in developing countries (Black et al 2013) Nevertheless studies connectingflood-exposure to the nutritional status of children have been rare and none of themconsidered the role of father education in preventing the health impacts of floodsamong children (Phalkey et al 2015) To the best of our knowledge no study to datehas investigated particularly the role of parental education on childrsquos nutritional statusin post-flood settings We only found one study which investigated the sole effect of theeducation of mothers not the education of fathers on malaria parasitemia in Sub-Saharanchildren (Siri 2014) Importantly no study has looked at the association of educationon child wasting let alone in the context of floods The investigation of child wastingis relevant as evidence suggest that stunting and wasting represent different processes ofundernutrition (Ricci amp Becker 1996) Recent evidence suggests that exposure to floodscan be associated to increases in both child stunting (Rodriguez-Llanes et al 2011) andchild wasting (Rodriguez-Llanes et al 2016) and thus both deserve attention in post-floodsettings

The large floods occurring in rural Odisha India in September 2008 produced massivedamage to agriculture water and sanitation communication networks and severedisruption to the normal functioning of the entire rural society at large (Governmentof Orissa 2008) As part of an integrated project to investigate the health social andeconomic impacts of disasters we carefully planned and conducted in September 2009a representative survey of children affected and non-affected by these floods in 265communities To get further insight into prevention strategies for the health impacts offloods which are absent in the literature (Bouzid Hooper amp Hunter 2013) we examined inthis study the effect of maternal and paternal education and other risk factors on stuntingand wasting in children from families living in flooded and non-flooded communities

MATERIALS amp METHODSDesign and sampleJagatsinghpur is a coastal district of the state of Odisha located within the Bay of BengalIndia The districtrsquos population is around one million 90 of which inhabit rural areas(The Census of India 2011) The region is located in a large and fertile flood plain crossedby large rivers which makes it very attractive for fishing activities livestock rearing andcrop farming However this land is also regularly affected by heavy monsoons which oftenproduce flooding In the last decade Jagatsinghpur has been hardly hit by five major floodsincluding coastal flooding associated to cyclone Paradip (05B) in 1999 followed by heavyrain floods in 2001 2003 2006 and the latest flood occurring prior our investigation whichtook place in mid-September 2008 (Government of Orissa 2008)

We conducted a population-based survey one year after the 2008 flooding in ruralJagatsinghpur district Odisha India A two-stage cluster survey was necessary to obtaina probability sample of children 6ndash59 months of age in 265 villages from four severelyflood-affected blocks of the district (Kujang Biridi Balikuda and Tirtol) Overall 122 ofthese were flooded and 143 non-flooded in September 2008 The percentage of households


Figure 1 Study site eligible villages and original sample of flooded and non-flooded villages in Jagats-inghpur district Odisha India Triangles represent flooded villages circles those non-flooded Size ofpolygons is proportionate to village size as measured by number of households (see map legend) Polygonsoverimpressed identified villages selected

flooded in each village was obtained for 2006 and 2008 events from OSDMA data (OdishaState Disaster Mitigation Authority 2009 Fig 1)

We used a two-stage cluster design to select our sample as our population of interest wasclustered in villages and the information on the population was scant (Groves et al 2009)In this study the Primary Sampling Units (PSUs) were the villages and the children werethe Secondary Sampling Unit (SSU) A 30 (PSU) by 30 (SSU) design which should providea probability sample of 900 children was fixed We initially enumerated the 265 villagesalong with each village population size projected for 2009 from census data (The Census ofIndia 2001) and subsequently 30 clusters from 29 villages were selected This first selectionwas done using Probability Proportionate to Size (PPS) sampling with replacement (ieone large village was selected twice) This method picks up villages randomly but thechances of selection are proportional to the size of the village with selection probabilitiesfavoring larger villages (Groves et al 2009) At the second stage a fixed number of children(ie 30) are selected by cluster independently of village size The result is that chancesare compensated with equal probabilities of selection of any eligible child listed in thesampling frame (Groves et al 2009) Importantly an updated list of eligible children wasobtained from the ICDS centers (1 ICDS center for every 1000 inhabitants) and validatedwith the ward members of each village A total 3671 eligible children were listed from 29


villages within a month prior training and piloting of the survey instrument Once the listscompiled it was detected that three flooded villages (Korana Jamphar and Raghunathpur)and two non-flooded (Muthapada Sureilo) did not have enough eligible children (ieless than 30) We created 5 new clusters by merging the list of children in each of these5 villages (n= 280) with those of the closest non-selected flooded or non-flooded villagerespectively of our list (Fig 1) Finally thirty children per cluster were randomly selected

Ethical approvalThis study was approved by the Community Health Ethics Committee Voluntary HealthAssociation of India New Delhi Persons eligible to participate in the study were notoffered any monetary incentive Written informed consent was obtained for every head ofhousehold visited In case the respondent was an illiterate we asked a literate person fromthe community to read out the consent form and explain it to the head of the family Wethen obtained the thumb impression of the respondent In those cases the person who readout the consent form also signed as a witness Research procedures were consistent withthe Declaration of Helsinki (1997) Interviews were administered after obtaining informedconsent The protocol was reviewed by a small group of scientists who had experienceworking with survivors of natural disasters and amended based on their recommendations

Data collection instruments and measuresOur survey instrument was adapted from the core one developed and approvedby a multidisciplinary consortium of researchers from the MICRODIS project Theinstrument development was based on interim literature reviews and follows the UNICEFconceptual framework on child malnutrition (United Nations Childrenrsquos Fund 1997) Thequestionnaire was validated prior to our study in research sites in India Indonesia thePhilippines Vietnam and the UK We collected background information at the householdlevel more specifically from mothers and fathers covering basic socio-demographiccharacteristics wealth child caring practices healthcare access maternal and paternaleducation income and credit practices water and sanitation food consumption patternsdemographics nutrition and health status data at the child level

To assess nutritional status using anthropometric indicators weight and heightlengthwere recorded Children were weighed without clothesWeight wasmeasured to the nearest100 g by trained research assistants using a beam balance (lt10 kg Raman Surgical CoDelhi-33 India) and an electronic balance for those children heavier than 10 kg For childrenyounger than 2 years of age length measurements were taken to the nearest millimeterin recumbent position with an infantometer (Narang Medical Delhi-110 028 India) Ifchildren were older than 2 years they were measured standing up with an adjustable boardcalibrated in millimeters Each research assistant measured and weighted each child twiceto minimize measurement errors and use the average value of both measurements to gainprecision These instruments were calibrated daily

Study questionnaires were administered by twelve experienced research assistants(Rodriguez-Llanes et al 2011) from the Voluntary Health Association of India (VHAI)who received specific training on anthropometry and interview procedures for this study


in late August 2009 The questionnaire was piloted in 12 households (6 in flooded villagesand 6 in non-flooded) and improved based on the inputs of the pilot exercise The studyquestionnaire was translated to the local language in Odisha (Oriya) and subsequentlyback translated into English by different professional translators and a researcher checkedthe level of agreement between both versions Duration of interviews ranged from 45 to60 min and all field work was completed between 6 and 24 September 2009

Study variablesAs our aim was to estimate the association effect of formal maternal and paternal educationon child undernutrition we excluded variables in our dataset which may have mediatedthis effect (Schisterman Cole amp Platt 2009) We supported our choice by examining thesevariables (eg caregiving practices food security and access to health care water andsanitation) framed within the UNICEF framework for child malnutrition (United NationsChildrenrsquos Fund 1997) Overall 51 variables were assessed in a previous study (Rodriguez-Llanes et al 2016) but for the purpose of this study we only analyzed distal determinantsof child undernutrition

Two outcomes were used in our study stunting (height-for-age) and wasting (weight-for-height) Stunting is an indicator of chronic malnutrition whereas wasting oftenevaluates acute nutritional stress at individual and population levels The new WHOstandard was used to calculate the z scores for these indicators Malnutrition was a binaryvariable indicating whether a children is malnourished z score le 2 (1) or not (0) at thetime of the interview

Overall 17 variables were examined as potential predictors The two fundamentalvariables of this study were the level of formal education attained by both childrenrsquos parentsmother and father Formal education was the only assessed in this study as we excludedparents with technical training or professional studies (Fig 2) We studied education usingthe same categories as recorded in the original questionnaire which follow benchmarksin the Indian educational system never attended school (0 years of schooling) completedelementary school (5 years of schooling) middle school (8 years) high school (12 years)and completed university studies (15 years or more) For mothers age at marriage age atfirst delivery and age at birth of the selected child were reported and analyzed in years thelater calculated by subtracting the age of each child (converted to years) to the motherrsquosage The fatherrsquos age at birth of the selected child was obtained by same calculationsChildrsquos sex was a binary variable The age of each child in months was obtained from birthcertificates and vaccination cards If these were not available local calendars were used Thebirthweight of each child was recorded in grams from birth certificates and for analyticalpurpose we expressed birthweight per 100 g The count of children younger than 5 yearsliving in a household were used in our analyses The principal means of livelihood weredichotomized as agriculture (taken as reference) grouping households dedicated to fishinglivestock rearing and crop farming versus non-agricultural for any other reported activityTwo religions were present in the study area Hinduism taken as the reference group andIslam The caste of the household was based on the household head and was grouped as ascheduled caste other backward class or general class (reference category) The general class


Figure 2 Flow diagram of sampling procedure sample obtained and analyses undertaken


is the higher caste status The scheduled caste is the social group historically subject to thehigher deprivation levels in the country Owned land was originally collected in acres butwas expressed in hectares and analyzed as a continuous variable Annual household incomewas recalculated per capita using data on household size and expressed per 1000 Indianrupees (INR) and modeled as a continuous predictor A household owning any livestockincluding chicken were modelled against those households not owning any Finally werecorded the exact number of persons residing in each household (household size) butwere dichotomized as more than four or otherwise to account for overcrowding as doneby a previous study (Semba et al 2008) The levels of occupation (employed unemployedor working as housewife) of mother and father were examined to better comprehend thelevel of gender empowerment within the household

Statistical methodsTwo expert data managers entered the data An external researcher (JMR) run exploratorydata analysis to identify errors in data entry and other implausible values (Day Fayers ampHarvey 1998) ENA for SMART software (version November 2008) was used to calculatenutritional indicators with the 2006 World Health Organization Standard (ENA 2008)ENA software is built up with specific functions for detection of outliers and impossiblevalues These were used to identify further problematic values among the calculatednutritional indicators Each potential error discovered was discussed and investigated todetermine where they originated and subsequently corrected

The main predictor in this study was mother and father education The original samplesize was determined based on requirements of a previous study (Rodriguez-Llanes et al2016) Regarding the main research question addressed by this report overall availablesample sizes were sufficient for subgroup analyses to detect prevalence ratios of two ormore with a 80 power (Sullivan Dean amp Soe 2009)

We examined the relationship between parental education and child nutritional statusseparately in repeatedly flooded and non-flooded cohorts A first consideration was not tomodel maternal and paternal education jointly but in separate analyses (Fig 2) Intuitivelyindividuals with similar education tend to get married more often and this was reflected byour data variance inflation factor (VIF) did show these two variables being highly collinearAs such eight models were fitted on this data (2 datasets times 2 outcomes times 2 predictorseach) These models are summarized in Fig 2

Bivariate and multiple adjusted logistic regression models with a quasi-binomialdistribution to control for overdispersion were fitted (Lumley 2010) For multivariatemodels we included only variables with p lt 01 in bivariate associations withundernutrition We first ran a full model with all variables having a p-value lower than01 a bit more conservative that some authors recommend (Vittinghoff et al 2005)On each full model the VIF on each predictor was calculated and predictors with thehighest VIF were sequentially eliminated until all remaining predictors had a VIF lowerthan four Backward selection was then applied to the remaining variables At each stepthe non-significant variable with the largest p-value was excluded to obtain a final modelincluding only significant variables at alpha level of 5 (ie p-value lt 005) As proposed in


recent literature (Siri 2014Muttarak amp Lutz 2014) we examined the impact of householdannual per capita income on the effects on paternal education to ensure that educationand income are independent contributors of child nutrition outcomes We did that in alleight models

Results were provided as prevalence ratios crude [PR] or adjusted [aPR] with 95confidence intervals Given the limits of our sample size interactions were not examinedAll tests were two-tailed with α= 005 All analyses were weighted Weights were calculatedas the inverse of the selection probabilities Statistical analyses were conducted in R (version302) (R Development Core Team 2008) with the survey package (Lumley 2010)

RESULTSStudy sampleA sample representative of the children population 6ndash59 months of age was obtained in 265villages of Jagatsinghpur district Our analyses excluded children with missing observationson relevant variables and from villages inundated in September 2008 but not in 2006Figure 2 provides further details on each stage of the sampling and exclusions beforestatistical analysis

Table 1 presents descriptive information for the variables analyzed in recurrentlyflooded communities (n= 299) and those non-flooded (n= 385) The illiteracy of motherswas less prevalent in flooded areas compared to non-flooded Whereas more motherscompleted middle school in flooded communities more completed university studies innon-flooded ones The percentage of fathers completing high school was 10 higher inflooded communities compared to non-flooded while those having university degrees wassimilar In general the flooded population was more educated relative to the non-floodedand fathers were more educated than mothers Almost 75 of the mothers attained at leastmiddle school amongst the flooded villages relative to 67 in the non-flooded For a similarcomparison amongst fathers the percentages were higher 89 and 812 respectively

Flooded households showed higher reliance on agriculture activities as per their higherpercentages of land and livestock owned which corresponded to higher proportionsdedicated to agricultural activities (418 in flooded vs 358 innon-flooded)Householdsof the highest cast were also more common in the sample of flooded villages whereasschedule and other backward castes were found more often in non-flooded communitiesUnemployment rates among fathers were 21 in non-flooded compared to 23 inrecurrently-flooded populations As many as 977 of all interviewed mothers worked ashousewives in both flooded and non-flooded communities

Correlates of child stuntingIn univariate analyses paternal and maternal education played an important role inreducing child stunting (Fig 3) Overall the protective effect of education on childstunting were substantial for most educated groups amongst flooded communities and inmen Compared to mothers never attending school a 3-fold (200) lower prevalence ofstunting was observed in most educated ones (ie those completing university studies)


Table 1 Prevalence of stunting wasting and baseline characteristics in rural flooded and non-flooded communities of Odisha India

Flooded (n= 299) Non-flooded (n= 385)

Variables n ormean (SE) n ormean (SE)

Stunting 299 304 (003) 385 290 (003)Wasting 299 515 (003) 385 203 (003)Maternal education

None 17 54 (001) 37 95 (002)Primary school 59 199 (003) 96 235 (002)Middle school 106 360 (003) 105 261 (002)High school 85 293 (003) 108 287 (003)College or more 32 93 (002) 39 122 (002)

Paternal educationNone 6 25 (001) 15 41 (001)Primary school 29 85 (002) 55 142 (002)Middle school 87 254 (003) 116 296 (003)High school 116 410 (003) 123 310 (003)College or more 61 226 (003) 76 212 (003)

Mean age at marriage of the mother years 299 218 (018) 385 215 (017)Mean age of the mother at first delivery years 299 237 (017) 385 234 (017)Mean age of the mother at birth of selected child years 299 261 (024) 385 259 (021)Mean age of the father at birth of selected child years 299 313 (027) 385 311 (027)Mean annual income per household capita 1000 rupees 299 75 (060) 385 73 (063)Means of livelihood linked to agriculture 299 418 (003) 385 358 (003)Mean land owned hectares 299 06 (005) 385 04 (004)Any livestock owned 184 647 (003) 209 552 (003)Religion

Hindu 248 899 (001) 357 927 (001)Muslim 51 101 (001) 28 73 (001)

CasteGeneral 83 332 (003) 68 205 (003)Other backward 100 324 (003) 183 419 (003)Scheduled caste 66 246 (003) 106 303 (003)No caste 50 98 (001) 28 73 (001)

No of individuals eating from same kitchen gt4 240 787 (003) 287 765 (002)Mean no of children under five eating from same kitchen 299 14 (004) 385 15 (005)Child female 140 450 (003) 175 444 (003)Mean birthweight 100 g 299 277 (025) 385 274 (021)Mean child age months 299 334 (108) 385 325 (097)

But generally the effects observed on paternal education were larger and stronger comparedto maternal education (Fig 3)

An additional hectare of land owned in flooded communities had a substantial protectiveeffect although non-significant (p= 007 detailed results in Table S1) but the effect was


Figure 3 Factors associated to child stunting in repeatedly flooded and non-flooded communities of rural coastal Odisha India Blue dots showthe prevalence ratios in repeatedly flooded communities Red dots in non-flooded Dotted lines show the relative difference in the univariate effectsbetween flooded and unflooded communities lowastlowastlowastplt 0001 lowastlowastplt 001 lowastplt 005 plt 01

lower in non-flooded A per capita household annual income improvement of 1000 rupeeswas also associated with a decrease of about 6 in child stunting but again only in floodedhouseholds (Fig 3 Table S1) Belonging to a backward caste was detrimental to childstunting but only in flooded communities and especially deleterious in the most deprivedcaste in India the scheduled caste with more than a twofold associated increase in theprevalence of child stunting

Multivariate analyses on child stuntingIn adjusted analyses the positive effect of maternal and paternal education on childnutrition weakened and remained statistically significant only for models on paternaleducation in flooded communities and marginally significant in non-flooded (Table 2) Inrepeatedly-flooded communities fathers with completed middle high school or universitystudies were consistently associated with 25- 21- and 27-fold lower prevalence of childstunting respectively In those non-flooded communities completedmiddle or high schoolby fathers conferred lower protection to their children with associated 19- (p= 0055) and18-fold (p= 0085) lower prevalence of stunting but none reached statistical significanceHaving completed college was comparably as protective in non-flooded areas as in flooded(28-fold reduction) and the effect statistically significant (p= 0015)

Per capita annual income was a consistent effect in flooded studied areas and acrossmodels on maternal and paternal education with 47ndash49 lower prevalence of stunting


Table 2 Multivariate logistic regressionmodels for maternal and paternal education and further risk factors associated to stunting in floodedand non-flooded children populations in rural Odisha India


aPR (95 CI) p-value aPR (95 CI) p-value

Models for maternal education ndash ndash ndash ndashMaternal education ndash ndash ndash ndash

No schooling 1 ndash 1 ndashPrimary school 0895 (0558 1434) 0645 1275 (0657 2475) 0473Middle school 0572 (0335 0978) 0042 0927 (0461 1866) 0832High school 0839 (0468 1504) 0556 0523 (0242 1127) 0099College or more 0598 (0228 1567) 0296 0717 (0263 1954) 0516

Child age (months) 1015 (1003 1027) 0014 1015 (1004 1026) 0008Caste ndash ndash ndash ndash

General 1 ndash NA NAOther backward 1654 (0864 3164) 0130 NA NAScheduled caste 2007 (1068 3770) 0031 NA NANo caste 1560 (0772 3151) 0216 NA NA

Per head annual income (per 1000 rupees) 0955 (0913 0998) 0043 NA NAModels for paternal educationPaternal education ndash ndash ndash ndash

No schooling 1 ndash 1 ndashPrimary school 0779 (0478 1268) 0315 0523 (0260 1052) 0070Middle school 0404 (0247 0661) lt0001 0530 (0278 1012) 0055High school 0487 (0300 0624) 0004 0563 (0293 1081) 0085College or more 0372 (0205 0791) 0047 0360 (0159 0816) 0015

Child age (months) 1013 (1001 1024) 0033 1012 (1001 1023) 0035Child birthweight (per 100 g) NA NA 0950 (0905 0997) 0037Per head annual income (per 1000 rupees) 0953 (0911 0996) 0032 NA NA

NotesNA results not available if variable not retained in multivariate model

for each yearly increase in 1000 rupees per capita at household level Schedule casteremained a relevant factor in multivariate analysis modeling the effect of maternaleducation but only in flooded communities (Table 2) The two final models fornon-flooded communities were also adjusted for the effect of income but no substantialchange in the coefficients were noted

Correlates of child wastingCrude associations showed an important role of parental education to reduce child wastingwith a preponderance of paternal over maternal effects of formal education and the largesteffects observed in non-flooded communities (Fig 4) An association between means oflivelihood (non-agriculture vs agriculture) and child wasting favoring households notrelying on agriculture activities was observed with a 47 lower prevalence of child wastingcompared to those dedicated to agricultural activities (Fig 4 detailed results in Table S2)


Figure 4 Factors associated to child wasting in repeatedly flooded and non-flooded communities of rural coastal Odisha India Blue dots showthe prevalence ratios in repeatedly flooded communities Red dots in non-flooded Dotted lines show the relative difference in the univariate effectsbetween flooded and unflooded communities lowastlowastplt 001 lowastplt 005 plt 01

Multivariate analyses on child wastingThe results for paternal education were different in adjusted models compared to thoseshowed by bivariate associations (Table 3 Table S2 and Fig 4)

In flooded communities the protective effects of different levels of paternal educationwere all significant (plt 0001) except for that of primary education (p= 0117) Howeverlarger effects were observed in the non-flooded communities surveyed in our study startingfrom primary education (29-fold lower prevalence p= 0019) middle (29 times lowerp= 0012) or high school (27-fold difference p= 0008) up to those with universitydegrees who were associated with two times lower prevalence of wasting among theirchildren though the association remained insignificant (p= 0081) In flooded householdsdedicated to activities other than agriculture a 50ndash51 lower prevalence of child wastingwas estimated (Table 3)

In modeling the effect of maternal education in flooded communities we found thatthe motherrsquos age at birth of the studied child was associated with an improved nutritionaloutcome of their children Each additional year was associated with a 34 lower prevalenceof wasting (p= 0005 Table 3) In flooded communities only and for themodel onmaternaleducation belonging to other backward castes showed to be deleterious on child wasting(Table 3) All four final models in this section were also adjusted for annual per headincome but none of those inclusions impacted our results


Table 3 Multivariate logistic regressionmodels for maternal and paternal education and further risk factors associated to wasting in floodedand non-flooded children populations in rural Odisha India





Mother age at birth of selected child (years) 0967 (0944 0989) 0005 NA NAMeans of livelihood (non-agricultural vs agricultural) 0660 (0526 0828) lt0001 NA NAChild age (months) NA NA 0975 (0956 0994) 0009Caste ndash ndash ndash ndash


Number of individuals eating from same kitchen ndash ndash ndash ndash2 4 NA NA 1 ndashgt4 NA NA 1870 (1059 3301) 0032

Models for paternal educationPaternal education ndash ndash ndash ndashNo schooling 1 ndash 1 ndashPrimary school 0777 (0567 1064) 0117 0348 (0144 0840) 0019Middle school 0571 (0432 0754) lt0001 0349 (0155 0787) 0012High school 0699 (0579 0844) lt0001 0366 (0175 0763) 0008College or more 0521 (0363 0750) lt0001 0490 (0220 1090) 0081

Means of livelihood (non-agricultural vs agricultural) 0667 (0526 0846) lt0001 NA NAChild age (months) NA NA 0977 (0958 0997) 0023


DISCUSSIONThis study identifies education as one key investment in reducing the health impactsof extreme events under climate change risks while promoting sustainable humandevelopment The most striking finding is that paternal education and not maternaleducation appeared to be the strongest predictor of lower child stunting and wasting inthese communities In non-flood settings large samples from Indonesia and Bangladeshstudied the effect of parental education on child stunting and comparable positive effectswere found for maternal and paternal education (Semba et al 2008) The effect of paternaleducation on child wasting and stunting was also found to be comparable to that foundin mothers in India and Vietnam (Moestue amp Huttly 2008) In flood settings however


these effects were not as clear as in available studies Hossain amp Kolsteren (2003) found noeffect of mother education on child wasting recovery four months after large floods inBangladesh Also in Bangladesh Del Ninno amp Lundberg (2005) found no effect of maternaleducation on child growth between three waves of anthropometric data collected within 15months after the floods None of these studies considered the effect of paternal education

In our specific setting we hypothesize that the observed differences in effects favoringpaternal education could be partly explained by the low economic independence of womenin our sample nearly 98 of mothers in both exposure groups were housewives andwe assumed that they had no revenues We acknowledge that this topic is complex andmore studies on this specific interaction should be conducted on datasets having largersamples than ours However in India evidence exist pointing to low maternal economicand physical autonomy as a contributor to child stunting (Shroff et al 2009 Imai et al2014)

A second observation is that the effect of paternal education on child stunting werelarger and statistically significant in flooded communities while the same estimated effectsin non-flooded communities were of smaller magnitude It looks as if the already positiveeffect of education might be boosted in a post-flood situation Several studies have shownthat educationmight substantially improve coping after disasters through avoiding incomeloss after a disaster (Garbero amp Muttarak 2013) diversification of economic activites (Vander Land amp Hummel 2013) or choosing sustainable mechanisms for coping prevail overshort-term views (Wamsler Brink amp Rantala 2012 Helgeson Dietz amp Hochrainer-Stigler2013) In a disaster situation decisions on allocation of the constraint resources availablecan be crucial More educated parents might be able to invest in coping strategies withlonger term benefits have more savings or simply work in professions which are lessaffected by floods Our results on child wasting showing that non-agricultural livelihoodswithstood better the shock caused by flooding reinforce this point Similar results on ahigher likelihood of post-disaster migration in low educated farmers are consistent withour views (Van der Land amp Hummel 2013) Crop farming livestock rearing and fishingwere the most affected activities according to reports from the Government of Odishaaround half the production of the 044 Million Kharif crops flooded were lost more than23 Million livestock flood-affected and above 6300 fishing boats with their nets andother equipment damaged (Government of Orissa 2008) Our findings are consistent withthis pattern of affectedness and reveal their mid-term consequences on child nutritionalhealth Young children of agricultural livelihoods were the most impacted regarding theprevalence of wasting one year after the floods which was around 50 higher than inother livelihoods Importantly no effect was observed for similar analyses conductedin non-flooded communities whether univariate or multivariate This reinforces thehypothesis that crop destruction and overall food insecurity is a very likely pathway tochild undernutrition among the flooded populations (Leaning amp Guha-Sapir 2013) Thereis much to do to in terms of mitigation strategies to reduce the initial impact of floodson the livelihoods of these vulnerable populations At the same time there is an urgentneed to reform and expand the relief response which was not commensurate to themagnitude and duration of the problem The basic needs of the affected were only covered


by the government on the first 15 days following the floods We need to ensure that theselivelihoods are satisfied on around a year after the floods and that we target most affectedlivelihoods Crops take months to be harvested since they are planted Animal stocks needtime to recover fully the same applies for repairing boats or buying new ones Withoutmore dedicated resources there is the risk of perpetuating poverty cycles

Another consistent result was the positive effect of income on child nutrition in flood-affected communities For example an increase in 5000 rupees per capita in yearly incomewithin households should be associated with 25 lower prevalence of child stuntingSustainable livelihood economic development is then a plausible strategy to reduce thenutritional burden of floods according to our data In contrast and again in flood-affectedlivelihoods only lower caste was associated with worst nutritional status for both wastingand stunting variables suggesting that caste is still in association with lower opportunitiesOur study findings show that these social determinants of different impacts get visible inan extreme situation such as a disaster but we were unable to offer an explanation on themechanism explaining this pattern in the data Qualitative research might plausibly bevery useful to investigate the reasons further Finally an important result was that mothersgiving birth later had a lower likelihood of having a wasted children Our model suggestthat by delaying five years the birth of a given child the associated prevalence of childwasting in case of floods would be on average 17 lower It is noteworthy to point that thisresult was independent from the effect of mother education adjusted for in this model

Taking all the above together education promotion in rural areas especially amongwomen would be extremely efficient In our study area about 30 of the mothers didnot completed middle school education which is mandatory in India There is also aneducation gap with men in our study area for whom this percentage was less than 20 a10 difference with women It is worth mentioning that looking across all models in ourstudy primary education (up to 11 years of age) was not a significant contributor to betterchild nutrition in seven out of eight models reported However providing three additionalyears of schooling (up to 14 years) had a significant positive impact on child nutritionMoreover the magnitude of effects reported for middle high and college education werequite comparable suggesting that at least in these rural communities middle education issufficient Our results are in agreement with the Indian policy on education More effortsare then needed to ensure that every Indian children receives the eight years of compulsoryformal education By promoting more education and adequate family planning (VanBraeckel et al 2012) women might naturally delay motherhood with potential benefitsto childrenrsquos nutritional health To ensure full benefits we hypothesized that mothereducation might lead to increased employment rates amongst women as this would alsocontribute to additional higher household income (with associated benefits shown by ourstudy) Notably it has been demonstrated that income allocation by mothers is moreefficient to improve the familyrsquos health including child nutritional status compared tofathers (Thomas 1990)


Strengths and limitations of the studyThere a few methodological considerations strengths and limitations that need to beaddressed here Firstly all the reported results were adjusted as an additional robustnesscheck by annual per capita household income As such all effects reported wereindependent of income and thus our results can be methodologically comparable tothose recently published (Muttarak amp Lutz 2014) Second we often found in the literaturematernal and paternal education jointly modelled (Semba et al 2008) However we foundthat these two variables were highly correlated in our data and enough evidence usingvariance inflation factor analysis which supported that they should be modelled separatelyFuture studies should be aware of this Third father education was highly correlated tocaste and motherrsquos age at birth of a child and this is the reason why models modelingfather education do not include these variables Strengths of this study include carefullyconsideration of income as well as having analyzed data on a non-flooded group whichallowed us to observe that some effects such as a motherrsquos age at birth income or castebecome important determinants only in the extreme circumstances of flooding Additionalpositive features of this study were its population-based design the consideration ofrecurrently flooded communities only coupled with the analysis of mid-term and morelong-term child undernutrition On the minus side our study was not based on a very largesample size which could have been missed the detection of smaller effects and providedtighter confidence intervals Similarly a larger sample size might have allowed us to analyzemost severe forms of malnutrition such as severe wasting and stunting In the aftermathof flooding in Bangladesh Choudhury amp Bhuiya (1993) have shown maternal education toplay a role in increasing severe forms of underweight Also as the findings presented arebased on cross-sectional data care is needed to not attribute causal relationships to theassociations showcased here Importantly it is difficult to generalize our results to othersettings such as urban slums as vulnerabilities might be different there However we areconfident that there might be similarities between our results and other rural areas aroundthe world in which subsistence farming is the norm With its limitations this is one of thefirst studies to look at child undernutrition in flood settings and study potential preventivefactors in short and long-term More studies are needed to solidify the evidence base forglobal action

CONCLUSIONSWe recommend strengthening the response to floods and targeting most vulnerableagricultural-dependent livelihoods which showed an associated 50 higher prevalence ofchild wasting Policies for relief should be reviewed to ensure a longer period of supportand proper take up by development programs which should give enough time and supportto these households relying on agricultural activities to fully recover

Education promotion in general and the strict fulfillment of schooling until thecompulsory 14 years of age is strongly recommended to protect childrsquos health in the face offuture flooding Policies effectively helping sustainable livelihood economic developmentand diversification of economic activities delayed motherhood which certainly includeeducation promotion are vividly praised


Given that global climate changes are set to increase flooding both in frequency andseverity which will further aggravate this situation we recommend urgent action to betaken (UNICEF Office of Research 2014)

ACKNOWLEDGEMENTSWe are grateful to JeanMacq Niko Speybroeck Sophie Vanwambeke and Philippe Donnenfor comments and guidance We are grateful to Pascaline Wallemacq for preparing themap

ADDITIONAL INFORMATION AND DECLARATIONS

FundingThe present research was funded by the European FP6 6th Framework Programme underThe MICRODIS ProjectmdashIntegrated Health Social and Economic Impacts of ExtremeEvents Evidence Methods and Tools (contract number GOCE-CT-2007-036877) Thefunders had no role in study design data collection and analysis decision to publish orpreparation of the manuscript

Grant DisclosuresThe following grant information was disclosed by the authorsEuropean FP6 6th Framework ProgrammeSocial and Economic Impacts of Extreme Events Evidence Methods and Tools GOCE-CT-2007-036877

Competing InterestsThe authors declare there are no competing interests

Author Contributionsbull Jose Manuel Rodriguez-Llanes conceived and designed the experiments analyzed thedata contributed reagentsmaterialsanalysis tools wrote the paper prepared figuresandor tables reviewed drafts of the paper had the research idea interpreted resultscontributed data to mapbull Shishir Ranjan-Dash conceived and designed the experiments performed theexperiments contributed reagentsmaterialsanalysis tools reviewed drafts of the papercontributed data to mapbull Alok Mukhopadhyay conceived and designed the experiments performed theexperiments reviewed drafts of the paperbull Debarati Guha-Sapir conceived and designed the experiments reviewed drafts of thepaper gave advice contributed to interpreting results

Human EthicsThe following information was supplied relating to ethical approvals (ie approving bodyand any reference numbers)

This study was approved by the CommunityHealth Ethics Committee VoluntaryHealthAssociation of India New Delhi An approval letter was issued by the Ethics Committee


Data AvailabilityThe following information was supplied regarding data availability

The raw data has been supplied as Data S1

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj1741supplemental-information

REFERENCESAhernM Kovats RSWilkinson P Few R Matthies F 2005 Global health impacts of

floods epidemiologic evidence Epidemiological Reviews 2736ndash46DOI 101093epirevmxi004

Alderman K Turner LR Tong S 2012 Floods and human health a systematic reviewEnvironment International 4737ndash47 DOI 101016jenvint201206003

Black RE Victora CGWalker SP Bhutta ZA Christian P De Onis M Ezzati MGrantham-McGregor S Katz J Martorell R Uauy R 2013Maternal and childundernutrition and overweight in low-income and middle-income countries Lancet382427ndash451 DOI 101016S0140-6736(13)60937-X

BouzidM Hooper L Hunter PR 2013 The effectiveness of public health interventionsto reduce the health impact of climate change a systematic review of systematicreviews PLoS ONE 8(4)e62041 DOI 101371journalpone0062041

Census of India 2001 New Delhi India the Registrar General amp Census CommissionerMinistry of Home Affairs Government of India Available at httpwwwcensusindiagovin (accessed 3 April 2009)

Census of India 2011 New Delhi India the Registrar General amp Census CommissionerMinistry of Home Affairs Government of India Available at httpwwwcensusindiagovin (accessed 9 September 2014)

Choudhury AY Bhuiya A 1993 Effects of biosocial variables on changes in nutritionalstatus of rural Bangladeshi children pre- and post-monsoon flooding Journal ofBiosocial Science 25(3)351ndash357

Day S Fayers P Harvey D 1998 Double data entry what value what price ControlledClinical Trials 19(1)15ndash24 DOI 101016S0197-2456(97)00096-2

Del Ninno C LundbergM 2005 Treading water The long-term impact of the1998 flood on nutrition in Bangladesh Economics and Human Biology 367ndash96DOI 101016jehb200412002

EM-DAT The OFDACRED International Disaster Database Available at httpwwwemdatbedatabase (accessed 15 November 2015)

Emergency Nutrition Assessment (ENA) Software for standardized monitoring andassessment of relief and transitions (SMART) version November 2008 Available athttpwwwnutrisurveyde ena enahtml (accessed November 2010)

Garbero A Muttarak R 2013 Impacts of the 2010 droughts and floods on communitywelfare in rural Thailand differential effects of village educational attainmentEcology and Society 18(4)27 DOI 105751ES-05871-180427


Government of Orissa 2008Memorandum on floods Orissa Special Relief Commis-sioner Orissa Revenue and Disaster Management Department

Groves RM Fowler FJ Couper MP Lepkowski JM Singer E Tourangeau R 2009Survey methodology Oxford Wiley-Blackwell

Helgeson JF Dietz S Hochrainer-Stigler S 2013 Vulnerability to weather disastersthe choice of coping strategies in rural Uganda Ecology and Society 18(2)2DOI 105751ES-05390-180202

Hossain SM Kolsteren P 2003 The 1998 flood in Bangladesh is different target-ing needed during emergencies and recovery to tackle malnutrition Disasters27172ndash184 DOI 1011111467-771700227

Imai KS Annim SK Kulkarni VS Gaiha R 2014Womenrsquos empowerment andprevalence of stunted and underweight children in Rural IndiaWorld Development6288ndash105 DOI 101016jworlddev201405001

Leaning J Guha-Sapir D 2013 Natural disasters armed conflict and public health NewEngland Journal of Medicine 369(19)1836ndash1842 DOI 101056NEJMra1109877

LindeboomM Llena-Nozal A Van der Klauuw B 2009 Parental education andchild health evidence from a schooling reform Journal of Health Economics28(1)109ndash131 DOI 101016jjhealeco200808003

Lowe D Ebi KL Forsberg B 2013 Factors increasing vulnerability to health effectsbefore during and after floods International Journal of Environmental Research andPublic Health 10(12)7015ndash7067 DOI 103390ijerph10127015

Lumley T 2010 Complex surveys a guide to analysis using R Hoboken John Wiley ampSons

Milman A Frongillo EA De Onis M Hwang JY 2005 Differential improvement amongcountries in child stunting is associated with long-term development and specificinterventions Journal of Nutrition 1351415ndash1422

Moestue H Huttly S 2008 Adult education and child nutrition the role of familyand community Journal of Epidemiology and Community Health 62(2)153ndash159DOI 101136jech2006058578

Muttarak R LutzW 2014 Is education a key to reducing vulnerability to naturaldisasters and hence unavoidable climate change Ecology and Society 19(1)42DOI 105751ES-06476-190142

Odisha State Disaster Mitigation Authority (OSDMA) 2009 Available at httpwwwosdmaorg (accessed 3 April 2009)

Phalkey RK Aranda-Jan C Marx S Houmlfle B Sauerborn R 2015 Systematic reviewof current efforts to quantify the impacts of climate change on undernutritionProceedings of the National Academy of Sciences of the United States of America112(33)E4522ndashE4529 DOI 101073pnas1409769112

RDevelopment Core Team 2008 R a language and environment for statistical comput-ing Vienna R Foundation for Statistical Computing Available at httpwwwR-projectorg

Ricci JA Becker S 1996 Risk factors for wasting and stunting among children in MetroCebu Philippines American Journal of Clinical Nutrition 63(6)966ndash975


Rodriguez-Llanes JM Ranjan-Dash S DegommeOMukhopadhyay A Guha-SapirD 2011 Child malnutrition and recurrent flooding in rural eastern India acommunity-based survey BMJ Open 1e000109 DOI 101136bmjopen-2011-000109

Rodriguez-Llanes JM Ranjan-Dash S Mukhopadhyay A Guha-Sapir D 2016 Flood-exposure is associated with higher prevalence of child undernutrition in rural easternIndia International Journal of Environmental Research and Public Health 13(2)210DOI 103390ijerph13020210

Schisterman E Cole S Platt R 2009 Overadjustment bias and unnecessary adjustmentin epidemiologic studies Epidemiology 20488ndash495DOI 101097EDE0b013e3181a819a1

Semba RD De Pee S Sun K Sari M Akhter N BloemMW 2008 Effect of parentalformal education on risk of child stunting in Indonesia and Bangladesh a cross-sectional study The Lancet 371322ndash328 DOI 101016S0140-6736(08)60169-5

Shroff M Adair L Suchindran C Bentley M 2009Maternal autonomy is inverselyrelated to child stunting in Andhra Pradesh IndiaMaternal Child Nutrition 564ndash74DOI 101111j1740-8709200800161x

Siri JG 2014 Independent associations of maternal education and household wealth withmalaria risk in children Ecology and Society 19(1)33 DOI 105751ES-06134-190133

Striessnig E LutzW Patt AG 2013 Effects of educational attainment on climate riskvulnerability Ecology and Society 18(1)16 DOI 105751ES-05252-180116

Sullivan KM Dean A SoeMM 2009 OpenEpi a web-based epidemiologic andstatistical calculator for public health Public Health Reports 124(3)471ndash474

Thomas D 1990 Intra household resource allocation an inferential approach Journal ofHuman Resources 25(4)635ndash664 DOI 102307145670

UNICEF Office of Research 2014 The challenges of climate change children on thefront line In Innocenti insight Florence UNICEF Office of Research

United Nations Childrenrsquos Fund 1997 The state of the worldrsquos children 1998 New YorkOxford University Press for UNICEF Available at httpwwwuniceforgnutritionfiles pub_sowc98_enpdf

Van Braeckel D TemmermanM Roelens K DegommeO 2012 Slowing populationgrowth for wellbeing and development The Lancet 38084ndash85DOI 101016S0140-6736(12)60902-7

Van der Land V Hummel D 2013 Vulnerability and the role of education in envi-ronmentally induced migration in Mali and Senegal Ecology and Society 18(4)14DOI 105751ES-05830-180414

Vittinghoff E Glidden DV Shiboski SC McCulloch CE 2005 Regression methodsin biostatistics linear logistic survival and repeated measures models New YorkSpringer

Wamsler C Brink E Rantala O 2012 Climate change adaptation and formal educa-tion the role of schooling for increasing societiesrsquo adaptive capacities in El Salvadorand Brazil Ecology and Society 17(2)2 DOI 105751es-04645-170202


48thWorldMedical Assembly 1997 Declaration of Helsinki recommendations guidingphysicians in biomedical research involving human subjects Journal of the AmericanMedical Association 277925ndash926 DOI 101001jama199703540350075038


impacts on improving children nutritional health in the long run especially underflooding Policies effectively helping sustainable livelihood economic development anddelayed motherhood are also recommended

Subjects Epidemiology Global Health Nutrition Pediatrics Public HealthKeywords Flood Climate change adaptation Wasting Stunting Father education Mothereducation Parental education Income Maternal autonomy Subsistence farming

INTRODUCTIONAmong all the disaster risks associated to a warming climate flooding has become themost frequent and since the nineties it has been affecting grosso modo 100 million peoplea year This is more than any other disaster type worldwide climate-related or not(EM-DAT 2015) Yet relative to their importance the health consequences of floodinghave been rarely investigated (Ahern et al 2005 Alderman Turner amp Tong 2012) Even ifat the time of writing an equivalent 378 (ie 28 billion people) of the current worldpopulation were affected by floods in the past 25 years little serious actions are being taken

Education has been recently stressed as one key to reduce the societal impacts ofclimate change (Striessnig Lutz amp Patt 2013) The study of educational attainment as apromotor of disaster resilience recently received substantive attention in a collection of11 papers (Muttarak amp Lutz 2014) The results of this special issue are a convincing stepforward These studies were undertaken in different countries and regions and targetedrelevant outcomes in different phases of the disaster cycle in populations exposed to diverseclimate-related hazards Overall they found that higher-educated groups avoided high-riskareas to settle were better prepared reacted more efficiently to early warnings and hadlessen impacts on health and social variables which indicated according to the authorsbetter coping and recovery in these groups (Muttarak amp Lutz 2014) For the particular caseof flooding the evidence on the role of education as a promotor of positive health outcomesis scant and contentious (Lowe Ebi amp Forsberg 2013) But what are the plausible pathwayslinking more education to health improvements Indeed formal education is crucialfor acquisition and processing of information (eg literacy) improvement of cognitiveabilities decision making and long-term planning and generally it leads to securing skilledjobs and ultimately higher income and better health (briefly reviewed in Striessnig Lutz ampPatt 2013 Lindeboom Llena-Nozal amp Van der Klauuw 2009)

The transgenerational effects of education on health have received substantial attentionAlthough the effect of maternal education on child stunting and general child healthhas been studied (Milman et al 2005 Lindeboom Llena-Nozal amp Van der Klauuw2009) the effect of paternal education on child health has been more rarely examined(Moestue amp Huttly 2008 Semba et al 2008) Assessing the extent to which the educationof the father affects a childrsquos health is an important step to understand the relativecontribution of fathers to the family wellbeing Considering the fatherrsquos education is alsoa premise to further study the synergies between father and mother education levels in thepromotion of childrsquos health (Semba et al 2008)












































Hindu 248 899 (001) 357 927 (001)Muslim 51 101 (001) 28 73 (001)































































































































































Hindu 248 899 (001) 357 927 (001)Muslim 51 101 (001) 28 73 (001)





















































































































looking upstream: enhancers of child nutritional status in …llanes,[email protected],...

Documents