7/29/2019 13_6_2007_1330_1342

1/13

7/29/2019 13_6_2007_1330_1342

2/13

Over the past 35 years, Egyptian demo-graphic and survival indicators have shownmarked improvement. For example, from1970 to 2005 life expectancy increasedfrom 52.1 years to 68.8 years, infant mortal-ity dropped from 157 to 35 deaths per 1000live births, and under-5 mortality droppedfrom 235 to 41 deaths per 1000 live births[1]. Despite these improvements in health

conditions there are still important impedi-ments to survival and development, espe-cially for children. One of the most serioushealth concerns is under-nutrition. Figure 1presents the trends in under-nutrition levelsin children in Egypt from 19922000 [1].While there are decreases in levels of childunder-nutrition, approximately 1 in 10 chil-dren (11%) under the age of 5 years wasunder-weight and approximately 1 in 5 chil-dren (21%) was under-height for age [2].

The nutritional level in children is avital component to their survival and de-

velopment in their early years. Low levels

of nutrition among children cause seriouslong- and short-term consequences in theirphysical and mental growth. Studies reporthigh levels of mortality among malnour-ished children [3]. Further, malnourishedchildren are more likely to have functionalimpairment in adult life [4] leading to areduction in productive life and thus af-fecting the overall economic productivityof the society [5]. For example, it is widelyaccepted that adults who survive malnutri-

tion as children are more likely to sufferfrom higher levels of chronic illness anddisability [6,7].

Studies on the nutritional status of chil-dren often accept the notion that it is deter-mined by a multiplicity of factors [79]. Inthis regard, several theoretical explanationsof malnutrition among children are foundin the literature [10]. These include fam-ily planning approaches [11,12], socioeco-nomic approaches [9,13] and the frameworkrecommended by UNICEF [14]. The former2 approaches emphasize a set of factors tothe exclusion of factors important to the

7/29/2019 13_6_2007_1330_1342

3/13

other approach. However, the UNICEFframework provides a holistic and prag-matic approach, in addition to addressingthe limitations in other approaches, to thestudy of nutrition of children in developingcountries. This framework follows Thetriple A approach (assessment, analysisand action) strategy to improve nutritionallevels in children [14]. Furthermore, theframework classifies the causes of malnutri-tion and death into 3 categories that account

for the complexity of the nutritional statusof children: basic causes at the societallevel, underlying causes at the household/family level, and immediate causes. Whilethis framework has strengths, one of themajor limitations is that many existing sec-ondary data sources do not allow research-ers to follow this framework in its entiretyin understanding child nutrition. The EgyptDemographic Health Survey [1], while notallowing as comprehensive an approachas would be ideal, provides variables thatcan be used within the basic and underly-ing causes identified in the framework.Furthermore, the present study incorporatesfamily planning, demographic and socio-economic approaches within the UNICEFframework.

Studies on child malnutrition in Egyptare often area-specific [15,16] and manystudies are limited to clinical approaches[17,18]. These studies have not exploredfully the influence of parental and socioeco-nomic characteristics within the UNICEFframework. While the earlier studies are

important contributions to the literature onEgyptian malnutrition, further research isneeded to understand the influence of theunderlying and basic causes determiningthe nutritional status of children in Egypt.The purpose of this paper therefore was toexplore basic and underlying factors deter-mining the nutritional status of children inEgypt.

The data for the present study come fromthe 2000 Egypt Demographic and HealthSurvey (EDHS) [1]. The 2000 EDHS wasthe sixth in the series of Demographic andHealth Surveys conducted in Egypt. Similarto the other surveys, data were collected onfertility, family planning, infant and childmortality, and maternal and child health

and nutrition. A nationally representativesample of 15 573 ever-married women aged1549 years were interviewed. This surveyincluded 2 questionnaires: a householdquestionnaire and an individual question-naire. The household questionnaire consist-ed of questions related to household socialand economic characteristics. The indi-vidual questionnaire included respondentsbackground (ever married women between1549 years), reproduction, contraceptiveknowledge and use, fertility preferencesand attitudes about family planning, preg-

nancy and breastfeeding, immunizationand health, schooling of children and childlabour, female genital mutilation, marriageand husbands background and womanswork and residence [1].

The primary objective of the EDHS2000 survey was to provide reliable esti-mates for fertility and child mortality forthe country and for 6 major administrativeregions. The methodology of the survey isdescribed in full in the EDHS report [1].Briefly, a 3-stage design was used to collect

a representative sample by which 17 521households were selected for the survey.From these households, the fully trainedfield staff interviewed 16 957 of the samplehouseholds, for a response rate of 99%. Allever-married women between 15 and 49years of age were eligible to participate inthe survey. As a quality control measure

7/29/2019 13_6_2007_1330_1342

4/13

10% of the households were selected forre-interview [1].

To assess the nutritional status of indi-vidual children, WHO recommends theuse of Z-score indicators of weight-for-age (WAZ) (under-weight), height-for-age(HAZ) (stunting) and weight-for-height(WHZ) (wasting). To compute the anthro-pometric indices, information on each in-

dividuals gender, age, weight, and heightare needed. WHZ and HAZ are the mostcommonly used indices for determiningnutritional status. The former is an indica-tor of wasting (i.e. thinness indicating acutemalnourishment) and the latter is an indi-cator of stunting (i.e. shortness indicatingchronic malnourishment). The third index,WAZ, is primarily a composite of WHZ andHAZ and is considered to represent acuteand chronic malnourishment. These indicespresent the long- and short-term prevalenceof malnutrition in children. In the ADHS2000 survey, heights for children youngerthan 24 months were measured lying on ameasuring board and standing height wasmeasured for older children. Weight datawere obtained using digital scales withan accuracy of 100 g [1]. For measuringchilds age a series of techniques were ap-plied in order to maintain accuracy. In ad-dition to asking mothers in what month andyear the child was born, the mothers werealso asked, How old was your child at his/her last birthday? In addition, interviewers

asked the mothers for birth cards or certifi-cates in cases where one was available andcross checked the responses.

The procedure for the computation ofthese measures and their interpretation arewell documented [1922]. In our study,the values of WAZ, WHZ and HAZ werecalculated using the Epi-Nut program pro-vided along with theEpi-Info, version 6.03.

This program transforms the internationalgrowth reference curves into a Z-score rep-resentation. These growth reference curveshave been used worldwide since 1978 toassess the nutritional status of children incross-sectional surveys. Smoothed nor-malized curves are fitted by polynomialregression and cubic spline techniques andthese curves are used to calculate all othernormalized Z-score values. The standarddeviations are defined separately for the up-

per and lower half of the skewed referencedistributions. It has been argued that the useof Z-score cut-off points provides interpre-tative guidance in that a known proportionof the reference population would be ex-pected to be below the cut-off point at anygiven age or height and for all indicators[19,20]. The commonly used conservativecut-off value of Z-score less than 2 SD isfollowed in our analysis.

Once the measures of malnutritionwere calculated for each of the 3 indices,measures of parental and socioeconomicconditions were identified. The basic causesat the societal level can be divided into2 types: potential resources and quantityand quality of actual resources. Potentialresources are understood in terms of po-litical, cultural, religious, economic andsocial subsystems present in the society andinclude conditions such as womens statusin the society [23]. The potential resources,in turn, influence the quantity and qualityof actual resources. The quantity and qual-ity of the resources available is shaped by

the human, economic, and organizationalmanner in which they are controlled [23].For the purposes of this paper, the basiccauses at the societal level were measuredin terms of place of residence, consanguin-ity between parents, parents level of educa-tion, mothers employment status, mothersheight, and childs sex.

7/29/2019 13_6_2007_1330_1342

5/13

Place of residence was identified asrural or urban. Consanguinity between theparents was measured as: no relation, firstdegree, second degree, or other blood rela-tive. Fathers education was measured as noformal education, primary school, second-ary school, or higher education. Motherseducation was measured as no formal edu-cation, primary school, secondary school,or higher education. Mothers employmentwas identified as working outside the home

or not working outside the home. Mothersheight was measured in centimetres as:< 150 cm, 150160 cm, or >160 cm. Thesex of the child was measured as male orfemale.

The second level identified in theUNICEF model considers underlying caus-es at the household/family level [6]. Centralto these underlying causes are the maternaland child care practices and include manyfactors linked to family planning activi-ties. In this paper, mothers age at birthof her last child, childs age, birth orderand birth interval were included. The ageof mother at birth was measured as < 19years of age, 1924 years, 2534 years or

35 years. The age of the child at the timeof data collection was measured in monthsas 011 months, 1223 months, 2435months, 3647 months or 4859 months.The birth order was identified as 1, 2, 3, 4 or5 and higher. Birth intervals were first born,023 months, 2435 months, 3647 monthsor 48 months.

A logistic regression technique was used

to estimate the odds of being malnourished.This technique permits control of the paren-tal and socioeconomic variables. To createthe dependent variable, the children whoseZ-scores were less than 2 SD were coded as1 and the children with Z-scores of 2 SD orhigher were coded as 0. The following wasthe basic model used in the analysis: Yij = a+ bXij+ cPj + mj+ eij; where: Yij: outcome

variable, Xij: individual control variables,Pj: variables at the community (PSU) level,mj: error of unobserved community vari-ables and ejj: error of unobserved individualvariables. The basic assumption is that mjis uncorrelated with the regressors; in otherwords, the above model considers the sam-ple design as well. Specifically, the modelpredicting the probability that the Z-scorevalue will fall below 2 SD (i.e. malnour-ished) takes the form: p(z = 1) = eY/(1 +

eY

).The predictor variables entered in the

regression equation are sets of dummyvariables. Thus, the results obtained werecompared with the reference category. Thepredictor variables used in the logistic re-gression model were: current place of resi-dence, sex of the child, mothers education,fathers education, birth spacing betweenthe child and the previous birth, birth order,age of mother at the time of the childsdelivery and age of the child at the timeof survey. The reference categories for thedifferent variables mentioned above were:living in a rural area, female child, maternalilliteracy, paternal illiteracy, birth spacing 160 cm tall. Children of mothers workingoutside the home had a lower prevalenceof stunting (17.76%) than those whosemothers who did not (18.82%). Stuntingwas higher among children born to mothersmarried to close relatives; approximately22% of children born to mothers married totheir first cousins (fathers or mothers side)were stunted compared to those born tomothers with no blood relation to their hus-bands. Stunting was higher in male children

(19.85%) than female children (17.42%).Higher levels of stunting were found inchildren of higher birth order (24.31% forbirth order 5 vs 17.26% birth order 1)and shorter birth intervals (23.44% for birthinterval < 23 months vs 16.7% birth interval

48 months). Children born to mothersaged < 19 years and 35 years showed ahigher prevalence of stunting than children

born to mothers in other age groups, 19.53%and 21.35% respectively. Just over 15% ofchildren aged 011 months at the time ofthe survey were stunted as were 23.44% ofchildren aged 1223 months.

While lower than the prevalence ofstunting, at 2.52% wasting was still greaterthan the expected 2.2%. There were no gen-erally clear patterns seen for stunting withthe different variables (Table 1). As regardsunder-nutrition, 4.06% of children under 5

years were under-nourished. The patternsseen were generally fairly similar to thoseof stunting (Table 1).

Although all 3 measures of under-nutri-tion were higher than would be expected,the highest prevalence was found in stunt-ing (chronic malnourishment) (HAZ);18.67% of the children under age 5 yearswere stunted. As this was the most preva-lent form of under-nutrition in Egypt, thesubsequent analysis focused on stuntingonly. Multivariate analysis was carried outto find the odds of stunting among children(Table 2). Since the birth interval variablewas calculated only for births of secondand higher orders, the multivariate analysisexcluded the first order births.

Several variables were found to havea significant influence on the prevalenceof stunting in Egypt. The odds of childrenbeing stunted in urban areas were 0.71 timeslower than for children in rural areas. Whilenot all categories of mothers educationwere statistically significant, they werein the expected direction. That is, as the

maternal educational level increased, theodds of children being stunted decreased.While no clear and significant pattern wasobserved with fathers education on stunt-ing, the odds of stunting were 0.76 timeslower among children whose fathers had atleast higher secondary education as com-pared to children born to fathers with noeducation. Mothers height was found to

7/29/2019 13_6_2007_1330_1342

7/13

7/29/2019 13_6_2007_1330_1342

8/13

have a significant influence on the odds ofstunting. Children born to mothers whoseheight was 150160 cm had 0.60 timeslower odds of stunting than children bornto mothers whose height was < 150 cmand odds were 0.41 times lower if moth-ers height was > 160 cm. Children whoseparents were first cousins had 1.21 higherodds of being stunted compared to childrenwhose parents were not blood relations;the odds of stunting were 1.22 times higher

if the parents were second cousins. Theodds of stunting were significantly loweramong low birth order children (3 or less)compared to children of birth order 5 andabove. The odds of stunting were higheramong children age 1223 months (1.79times) and 2435 months (1.28 times) thanchildren aged < 12 months. Birth intervalshad a significant influence on stunting; the

odds of being stunted declined as the birthinterval increased. Other variables foundnot to be significantly associated with theodds of stunting were: mothers workingstatus, male sex and age of mother at timeof birth (Table 2).

Measuring the weight and height of thechild actually measures much more than asingle child, these also measure the futureof a country. Child health has a prominentrole in shaping and defining the structureof a society. It shapes the quality of futurehuman capital, helps population stabiliza-tion and furthers future economic growth,among other factors. In the case of Egypt,there have been marked improvements over

7/29/2019 13_6_2007_1330_1342

9/13

7/29/2019 13_6_2007_1330_1342

10/13

the past 10 years in the nutritional statusof children but much work remains to bedone. In order to further decrease levelsof childhood malnutrition in Egypt, policyframeworks must be established that incor-porate short-term, medium-term and long-term strategies to solve nutritional problems[24]. The intervention strategies should becomprehensive, culturally sensitive and, asmalnutrition is a public health concern, ad-dressed at various levels of government.

Based on the findings of this research,policy development must take into consid-eration the ruralurban divide in malnutri-tion. It is clear that children raised in ruralareas are at greater risk for under-nutrition.Our results support the findings of priorstudies that have also described the urbanrural differences in health in Egypt. Peopleliving in urban areas are provided withbetter access to health services, educationand other social support systems which areeither not available or not easily accessibleto residents in rural areas. For example,studies have shown that immunization ratesare higher in urban areas as compared to ru-ral areas of Egypt [25]. Programmes shouldthus be developed to analyse and implementappropriate strategies to address rural andurban child malnutrition. For example, inEgypt most nutritionists are based in urbanhospitals while cases of malnutrition inrural areas are unlikely to seek care in suchhospitals [26]. Improvements in access tocare in rural areas together with the intro-duction of awareness programmes will have

a significant impact on ameliorating theexisting nutritional conditions in children.

Programmes should be developed thattarget higher risk groups such as youngchildren and higher birth order children.Further research and needs assessments arerequired to examine this situation in order todesign intervention programmes. The factthat risks are different at different ages sug-

gests the need for age-specific interventionsto address the nutritional needs of children.Additionally, family planning conditionssuch as birth interval, which was found tobe very important in our study, need to beconsidered. This finding supports the ideathat rearing children requires a large amountof resources and attention. When childrenare born closely together, the demands onresources such as mothers time, food, andother resources may be greater than the

family can provide. Our findings thereforesuggest that greater spacing of childrencould help to address under-nutrition byrelieving competition and exhaustion ofavailable resources.

While socioeconomic conditions cer-tainly have an effect on the nutritional statusof children, there is growing recognitionthat genetic factors must be consideredas well. There have been calls for the in-clusion of genetic characteristics in thestudy of child health and nutrition [27,28]but genetic components have not been ad-equately examined in many studies [29].The inclusion of 2 genetic factors in thisstudy is an important contribution to theliterature on childhood nutritional status.The 2 genetic characteristics (mothersheight and parental consanguinity) emergedas significant factors influencing stuntingof children in Egypt. Consanguinity in apopulation depends on several factors suchas demographic, social and religious normsand values. Many studies have shown con-flicting results on the impact of consan-

guineous marriages on childhood health[3032]. Children born to consanguineousparents are at an increased risk of autosomalrecessive disorders, multifactorial diseases[30] and early postnatal mortality [33,34].These disease processes can produce disor-ders similar to nutrient deficiency and mayresult in the appearance of stunting [24].Although there are conflicting opinions,

7/29/2019 13_6_2007_1330_1342

11/13

nutritional studies have seldom includedthis phenomenon. More studies are thusneeded in this area to explore further theinfluence of genetic characteristics on earlychildhood nutrition in Egypt and elsewhere.From a policy implementation perspective,the issue of consanguinity will require cul-tural sensitivity to the issue [24]. To addressconsanguinity, educational programmescould be instigated to create awareness,particularly among young unmarried males

and females, about the potential risks as-sociated with consanguineous marriage.Additionally, genetic counselling and riskassessment for consanguineous coupleswould assist families in making familyplanning decisions that could have an im-pact on the nutritional and health status oftheir children.

The results of our study show the impor-tance of the UNICEF model incorporatingparental and socioeconomic characteristicsin understanding the prevalence of under-nutrition, especially stunting, in Egypt.The framework classifies the causes ofmalnutrition to account for the complexityof the nutritional status of children. At thebasic level, rural conditions and the statusof women in society have an important con-tribution to the nutritional status of children.Specifically, children in rural areas are atgreater risk of stunting. Regarding womensstatus in society, education and the practiceof consanguinity contributed to the child-hood malnutrition. Family planning factorsas underlying causes at the household/fam-

ily level were also found to be importantaspects of the UNICEF model. Specifically,birth order and birth interval were importantfactors and relate to inadequate maternaland childcare practices. Further researchis needed to explore the influence of ad-ditional basic level causes and underlyingcauses at the family/household level as well

as inclusion of immediate causes such asthe presence of disease and dietary intakein the child.

These factors should be considered inthe development of any policies or pro-grammes aimed at alleviating the problemof under-nutrition. Among all factors, pa-rental characteristics, mothers character-istics in particular, were found to have asignificant impact on determining the nu-tritional status of children. Empowerment

of women through increasing educationallevels, choice in marriage, family planning,and other activities that improve the statusof women in society must be a priority inprogramme and policy development aimedat addressing the factors associated withunder-nutrition in children in Egypt.

Hameida and Billot argue that to meas-ure the weight and height of a child is tomeasure his or her health [35]. In thepresent study we assessed the factors thatcontribute to the health condition of thechild. More specifically, we consideredbasic and underlying conditions that affectthe nutritional status of children in Egypt.The study results further strengthen ourunderstanding about the nutritional status ofchildren in Egypt. Although, at the nationallevel, the prevalence of under-weight andwasting has decreased since 1995, theseestimates are still higher than would be de-sired. Most alarmingly, the high prevalenceof stunting signifies a public health prob-lem; the anthropometric measures showestimates of stunting above international

standards. Stunting is so prevalent that al-most one child in every five children underthe age of 5 years in Egypt is stunted. Thedocumented decrease in childhood stuntingin Egypt is promising but more work mustbe done to address the issues of childhoodmalnutrition if the gains are to be sustainedand further progress achieved.

7/29/2019 13_6_2007_1330_1342

12/13

7/29/2019 13_6_2007_1330_1342

13/13