Skip to main content
NCBI home page
Bulletin of the World Health Organization logoLink to Bulletin of the World Health Organization
. 2008 Feb 19;86(4):282–291. doi: 10.2471/BLT.07.044800

Socioeconomic inequality in malnutrition in developing countries

Inégalités socioéconomiques face à la malnutrition dans les pays en développement

أثر الجور الاجتماعي والاقتصادي في الإصابة بسوء التغذية في البلدان النامية

Ellen Van de Poel a,✉, Ahmad Reza Hosseinpoor b, Niko Speybroeck c, Tom Van Ourti a, Jeanette Vega b
PMCID: PMC2647414  PMID: 18438517

Abstract

Objective

The objectives of this study were to report on socioeconomic inequality in childhood malnutrition in the developing world, to provide evidence for an association between socioeconomic inequality and the average level of malnutrition, and to draw attention to different patterns of socioeconomic inequality in malnutrition.

Methods

Both stunting and wasting were measured using new WHO child growth standards. Socioeconomic status was estimated by principal component analysis using a set of household assets and living conditions. Socioeconomic inequality was measured using an alternative concentration index that avoids problems with dependence on the mean level of malnutrition.

Findings

In almost all countries investigated, stunting and wasting disproportionately affected the poor. However, socioeconomic inequality in wasting was limited and was not significant in about one third of countries. After correcting for the concentration index’s dependence on mean malnutrition, there was no clear association between average stunting and socioeconomic inequality. The latter showed different patterns, which were termed mass deprivation, queuing and exclusion. Although average levels of malnutrition were higher with the new WHO reference standards, estimates of socioeconomic inequality were largely unaffected by changing the growth standards.

Conclusion

Socioeconomic inequality in childhood malnutrition existed throughout the developing world, and was not related to the average malnutrition rate. Failure to tackle this inequality is a cause of social injustice. Moreover, reducing the overall rate of malnutrition does not necessarily lead to a reduction in inequality. Policies should, therefore, take into account the distribution of childhood malnutrition across all socioeconomic groups.

Introduction

Epidemiological evidence points to a small set of primary causes of child mortality that are the main killers of children aged less than 5 years: pneumonia, diarrhoea, low birth weight, asphyxia and, in some parts of the world, HIV and malaria. Malnutrition is the underlying cause of one out of every two such deaths.1,2 The evidence also shows that child death and malnutrition are not equally distributed throughout the world. They cluster in sub-Saharan Africa and south Asia, and in poor communities within these regions.3,4 Disparities in health outcomes between the poor and the rich are increasingly attracting attention from researchers and policy-makers, thereby fostering a substantial growth in the literature on health equity.5–8 “Socioeconomic inequality” in malnutrition refers to the degree to which childhood malnutrition rates differ between more and less socially and economically advantaged groups. This is different from “pure inequality”, which takes into account all factors influencing childhood malnutrition. The available literature documenting socioeconomic inequality in malnutrition focuses mainly on individual countries or regions.9–14 At a more global level, Wagstaff and Watanabe15 provided evidence on socioeconomic inequality in malnutrition across 20 developing countries. Other relevant cross-country studies include those of Pradhan et al.,16 who describe total inequality, and Smith et al.,17 who describe inequalities between urban and rural populations. The latter two studies, however, provide no evidence on socioeconomic inequality within developing countries.

This paper contributes to the literature in several ways. First, it updates and enlarges the evidence base on average malnutrition and socioeconomic inequality in malnutrition using the most recent Demographic and Health Survey (DHS) data from 47 developing countries. The inclusion of such a large number of countries makes it possible to obtain insights into the regional clustering of poor–rich malnutrition disparities in the developing world and into the association between the average level of malnutrition and socioeconomic inequality. Given the focus on average rates of malnutrition in international development targets, it is of interest to establish how countries compare in terms of average rates of malnutrition and inequality in malnutrition. In addition to quantifying the degree of socioeconomic inequality using a single index, this paper also illustrates the different patterns found for the distribution of malnutrition across socioeconomic groups.

Second, in this paper, childhood malnutrition is measured using the new growth standards that have recently been released by WHO.18 The new standards are based on children from Brazil, Ghana, India, Norway, Oman and the United States of America, and adopt a fundamentally prescriptive approach that is designed to describe how all children should grow, rather than merely how they actually grew in a single reference population at a specified time.19 For example, the new reference population only includes children from study sites where at least 20% of women were willing to follow breastfeeding recommendations. To the best of our knowledge, this is the first study that presents estimates of malnutrition based on these new standards in a large set of countries. To check the sensitivity of the results to this change in reference group, the analysis was also carried out using the older United States National Center for Health Statistics (NCHS) reference population.20

Finally, in this paper, socioeconomic inequality in malnutrition is measured using the concentration index, which takes into account inequality across the entire socioeconomic distribution. Usually, when it is applied to binary indicators, such as mortality and stunting, the concentration index depends on the mean of the indicator. This would impede cross-country comparisons because there are substantial differences in means between locations. To avoid this problem, we use an alternative but related index recently introduced by Erreygers.21

Methods

Data

The data used came from all 47 DHSs that contained information on the nutritional status of children aged up to 5 years. The data represent countries from four regions: 26 in sub-Saharan Africa, seven in the eastern Mediterranean, five in south and south-east Asia, and nine in Latin America and the Caribbean. Table 1 shows the countries and the characteristics of the data sets used.

Table 1. Characteristics of Demographic and Health Survey (DHS) data sets.

Country Year of 
survey Sample 
size Country Year of 
survey Sample 
size
Sub-Saharan Africa
 Eastern Mediterranean
Benin 2001 3 842 Armenia 2000 1 517
Burkina Faso 2003 8 142 Egypt 2000 10 296
Cameroon 2004 3 168 Kazakhstan 1999 566
Central African Republica 1994/1995 2 297 Kyrgyzstana 1997 971
Chad 2004 4 414 Morocco 2003/2004 5 356
Comorosa 1996 921 Turkey 1998 2 782
Côte d’Ivoire 1998/1999 1 477 Uzbekistan
 1996
 954
Ethiopia 2000 2 833 South and south-east Asia
Gabon 2000 3 482 Bangladesh 2004 5 911
Ghana 2003 3 094 Cambodia 2000 3 522
Guinea 1999 2 961 Indiaa 1998/1999 24 989
Kenya 2003 4 719 Nepal 2001 6 163
Madagascar 2003/2004 2 908 Pakistan
 1990/1991
 4 079
Malawi 2000 9 162 Latin America and the Caribbean
Mali 2001 9 382 Bolivia 2003 9 134
Mauritania 2000/2001 3 306 Brazil 1996 4 056
Mozambique 2003 3 808 Colombia 2005 12 393
Namibia 2000 2 925 Dominican Republic 2002 9 288
Nigera 1998 3 914 Guatemala 1998/1999 3 879
Nigeria 2003 4 293 Haiti 2000 5 510
Rwanda 2000 6 038 Nicaragua 2001 5 875
Togoa 1998 3 443 Paraguay 1990 3 614
Uganda 2000/2001 5 145 Peru 2000 11 585
United Republic 
of Tanzania 2004 7 132
Zambia 2001/2002 1 932
Zimbabwe 1999 2 632
Open in a new tab

a Births in the 3 years preceding the survey instead of the usual 5 years.

Analysis

Anthropometric data on the height-for-age and weight-for-height of children were used to quantify chronic and acute malnutrition, respectively. A small height-for-age reflects the slowing of skeletal growth, and is considered to be a reliable indicator of long-standing malnutrition in childhood. Low weight-for-height, on the other hand, indicates a deficit in tissue and fat mass, and this measure is more sensitive to temporary food shortages and episodes of illness. A low weight-for-age is also used in the literature to indicate malnutrition, but it is not used here as it does not discriminate well between temporary and more permanent malnutrition.9,20,22

A child was considered stunted or wasted if his or her height-for-age or weight-for-height, respectively, was two standard deviations or more below the median for the reference population.9,16 We used these crude binary indicators of stunting and wasting because their average values are much easier to interpret intuitively than continuous height-for-age and weight-for-age z-scores, and they, therefore, facilitate the comparison of stunting and wasting rates across socioeconomic groups and between countries.

This paper used the new WHO child growth standards that were released in April 2006.18 The robustness of the paper’s results against this change from the NCHS growth standards was also checked.20

An indicator of socioeconomic status was developed using principal component analysis.23 This indicator combined information on a set of household assets and living conditions: the ownership of a car, phone, television, radio, refrigerator, bicycle and motorcycle; the availability of electricity, clean water and a toilet; and the material used to construct the wall, roof and floor of the household dwelling.

Socioeconomic inequality in stunting and wasting were calculated by means of a recently proposed generalization – introduced by Erreygers21 (see also Van de Poel et al.24 for an application) – of the traditional concentration index (C), which was proposed by Wagstaff et al.25 This generalization overcomes several of the methodological shortcomings of the traditional concentration index while preserving its main characteristics: (i) negative values imply that malnutrition is more concentrated among poorer children, (ii) if all children, irrespective of their socioeconomic status, suffer equally from malnutrition, the concentration index would equal zero, and (iii) transferring malnutrition from a richer to a poorer individual reduces the concentration index. Of particular importance for this paper, it is worth mentioning that the generalization avoids dependence on the mean for the binary indicator (Wagstaff discussed a related issue for the bounds of the concentration index).26 Not correcting for this dependence on the mean would impede cross-country comparisons because there are substantial differences in means between locations. In addition, it would result in a predetermined association between the average level of malnutrition and socioeconomic inequality.

Since the DHSs rely on multistage sampling procedures, all estimates take account of sampling weights, and statistical inference is adjusted for clustering at the level of the primary sampling unit. The statistical inference for the index recently proposed by Erreygers was based on an adapted version of the convenient regression approach.27,28

Results

Table 2 shows socioeconomic inequality in stunting. In almost all countries, stunting disproportionately affected the poor. The concentration indices (based on WHO child growth standards and calculated as suggested by Erreygers)21 were significant in all countries, except Madagascar, and ranged from –0.0005 in Madagascar to –0.42 in Guatemala. Socioeconomic inequality in stunting appeared largest in the Latin America and the Caribbean region, where the median concentration index equalled –0.22.

Table 2. Estimated stunting rates in children aged less than 5 years according to socioeconomic status.

Country Prevalence of stunting by socioeconomic 
status quintilea (%)
 Average 
stuntingb (%)
 Concentration 
indexb,c
Q1 Q2 Q3 Q4 Q5 WHO NCHS WHO NCHS
Sub-Saharan Africa
Benin 43.78 45.38 39.98 34.96 27.35 38.61 30.37 –0.15 –0.13
Burkina Faso 48.44 46.96 46.49 40.20 27.45 42.98 38.56 –0.15 –0.15
Cameroon 44.19 43.42 38.85 31.25 19.20 36.49 31.68 –0.21 –0.21
Central African Republic 47.26 41.80 39.89 42.03 33.22 39.84 33.65 –0.11 –0.12
Chad 48.62 44.84 46.07 39.43 33.92 44.16 40.95 –0.09 –0.09
Comoros 46.11 47.08 41.45 37.97 26.47 40.53 33.77 –0.15 –0.19
Côte d’Ivoire 38.66 29.41 31.07 26.10 19.28 31.26 25.17 –0.17 –0.17
Ethiopia 60.94 55.04 58.23 54.07 42.27 56.91 51.22 –0.09 –0.10
Gabon 43.46 35.53 26.44 18.17 18.17 26.03 20.65 –0.22 –0.20
Ghana 45.11 38.27 40.42 30.42 20.01 35.62 29.43 –0.19 –0.19
Guinea 39.08 38.87 35.50 32.42 24.95 34.44 26.07 –0.13 –0.11
Kenya 43.18 39.34 35.48 27.98 22.87 35.90 30.56 –0.17 –0.16
Madagascar 53.90 54.72 59.96 58.15 50.51 56.06 48.34 –0.00d –0.01d
Malawi 60.64 59.59 52.80 57.79 39.32 54.08 49.02 –0.14 –0.14
Mali 48.79 49.60 45.10 42.40 28.43 41.78 37.57 –0.17 –0.17
Mauritania 45.05 41.47 40.69 32.80 31.65 39.25 34.50 –0.14 –0.16
Mozambique 55.79 53.08 53.84 43.45 34.70 51.50 46.16 –0.11 –0.14
Namibia 33.10 31.68 23.87 18.45 25.00 28.07 22.64 –0.13 –0.09
Niger 50.81 49.09 46.26 49.30 36.53 47.05 41.08 –0.08 –0.09
Nigeria 54.30 50.13 49.55 36.33 25.20 43.19 38.41 –0.25 –0.25
Rwanda 52.34 51.60 51.52 47.00 31.88 47.21 42.37 –0.14 –0.15
Togo 37.45 34.25 30.05 25.88 19.03 30.37 21.72 –0.16 –0.14
Uganda 45.84 46.75 49.46 42.79 29.00 44.50 38.61 –0.07 –0.08
United Republic of Tanzania 48.17 48.22 48.22 44.22 23.91 43.63 37.05 –0.15 –0.16
Zambia 59.53 58.41 58.33 49.88 40.59 53.21 46.15 –0.17 –0.18
Zimbabwe 37.37 34.65 32.33 29.87 23.45 31.48 26.45 –0.11 –0.12
Median 46.69 46.06 43.28 38.70 27.40 41.15 35.77 –0.15 –0.15
South and south-east Asia
Bangladesh 58.19 55.89 53.32 43.03 30.26 49.85 43.02 –0.20 –0.20
Cambodia 54.32 52.78 48.60 43.51 39.86 48.47 44.29 –0.15 –0.16
India 56.43 53.35 49.02 45.54 41.56 49.68 43.75 –0.13 –0.13
Nepal 63.76 63.40 58.92 47.08 42.01 56.46 50.51 –0.19 –0.18
Pakistan 61.91 62.94 53.58 49.13 35.98 54.12 49.59 –0.20 –0.24
Median 58.19 55.89 53.32 45.54 39.86 49.85 44.29 –0.19 –0.16
Eastern Mediterranean
Armenia 25.08 26.01 14.88 14.01 12.45 18.36 13.00 –0.12 –0.09
Egypt 31.80 26.41 22.69 19.23 15.18 24.00 18.66 –0.13 –0.12
Kazakhstan 17.81 14.91 9.29 9.40 6.32 13.93 9.75 –0.10 –0.10
Kyrgyzstan 41.40 37.66 24.36 28.64 18.88 32.89 24.84 –0.18 –0.17
Morocco 34.87 26.06 20.07 16.68 16.02 23.28 18.18 –0.18 –0.17
Turkey 34.25 23.52 17.48 9.50 5.01 19.04 16.01 –0.24 –0.22
Uzbekistan 41.12 38.35 32.21 33.77 36.00 37.46 31.28 –0.07 –0.09
Median 34.25 26.06 20.07 16.68 15.18 23.28 18.18 –0.13 –0.13
Latin America and the Caribbean
Bolivia 48.50 39.71 29.68 22.87 14.29 32.43 26.38 –0.31 –0.29
Brazil 29.46 13.25 7.61 5.41 5.42 13.42 10.46 –0.22 –0.19
Colombia 25.14 17.19 13.89 10.59 6.39 15.70 11.52 –0.15 –0.13
Dominican Republic 21.11 13.51 12.44 8.28 7.45 11.76 8.85 –0.12 –0.10
Guatemala 68.45 67.75 64.23 43.06 25.46 52.80 46.37 –0.42 –0.42
Haiti 38.01 33.83 29.97 21.65 11.74 27.10 21.93 –0.22 –0.19
Nicaragua 42.16 31.73 22.14 12.05 9.46 24.67 20.13 –0.30 –0.27
Paraguay 28.52 24.60 20.84 11.00 7.17 18.20 13.92 –0.20 –0.18
Peru 54.91 43.00 24.91 17.00 14.36 31.29 25.42 –0.41 –0.38
Median 38.01 31.73 22.14 12.05 9.46 24.67 20.13 –0.22 –0.23
Open in a new tab

a Based upon WHO Multicentre Growth Reference Study (MGRS).
b Derived using WHO and United States National Center for Health Statistics (NCHS) growth standards.
c Calculated as suggested by Erreygers.21
d Social inequality is not significant at the 10% level.

The results on wasting are presented in Table 3. Wasting was generally more concentrated among the poor, but socioeconomic inequality was much smaller than for stunting. For about one third of countries, socioeconomic inequality was not significant. The median concentration index was largest in south and south-east Asia (–0.05 based on WHO child growth standards).

Table 3. Estimated wasting rates in children aged less than 5 years according to socioeconomic status.

Country Prevalence of wasting by socioeconomic 
status quintilea (%)
 Average 
wastingb (%)
 Concentration 
indexb,c
Q1 Q2 Q3 Q4 Q5 WHO NCHS WHO NCHS
Sub-Saharan Africa
Benin 12.09 12.06 8.42 7.94 5.76 9.33 7.55 –0.06 –0.04
Burkina Faso 22.01 23.04 23.22 21.27 15.50 21.48 18.72 –0.04 –0.02d
Cameroon 8.24 8.46 5.86 4.00 2.93 6.23 5.28 –0.06 –0.05
Central African Republic 10.64 10.79 10.53 8.55 7.44 9.25 7.18 –0.03 –0.03
Chad 17.69 14.89 15.90 16.77 15.88 16.09 13.53 0.00d 0.00d
Comoros 15.52 13.78 10.36 5.91 8.43 11.00 8.40 –0.08 –0.06
Côte d’Ivoire 7.80 8.25 5.66 5.06 4.25 6.85 7.80 –0.02d –0.04
Ethiopia 13.11 13.51 13.52 12.19 7.10 12.70 10.71 –0.02 –0.04
Gabon 4.35 3.02 5.33 5.17 3.27 4.26 2.83 0.00d –0.01d
Ghana 8.57 7.90 8.67 10.20 8.15 8.70 7.12 0.00d 0.00d
Guinea 12.38 10.02 10.48 8.34 8.27 9.92 9.17 –0.04 –0.02
Kenya 8.70 5.35 4.80 3.65 7.59 6.23 5.62 –0.05 –0.05
Madagascar 11.83 11.40 9.17 8.95 7.19 10.04 7.75 –0.04 –0.04
Malawi 8.71 7.32 6.92 6.62 5.76 7.02 5.52 –0.02 –0.02
Mali 12.68 15.49 14.24 13.26 9.49 12.91 10.65 –0.04 –0.04
Mauritania 18.25 16.26 15.20 12.04 12.38 15.27 13.40 –0.06 –0.07
Mozambique 8.44 5.88 5.99 5.39 4.70 6.55 4.60 –0.03 –0.02d
Namibia 13.76 8.61 7.71 6.53 9.14 9.85 8.91 –0.08 –0.04
Niger 30.78 27.24 27.02 25.25 14.98 25.66 20.63 –0.08 –0.09
Nigeria 12.41 13.76 9.98 10.98 9.11 11.34 9.48 –0.03 –0.03
Rwanda 9.11 10.52 8.69 8.14 7.66 8.88 6.85 –0.02 –0.02
Togo 13.86 19.59 13.48 12.17 8.57 13.98 12.42 –0.06 –0.04
Uganda 5.37 5.15 5.99 4.60 3.50 5.11 4.04 –0.01d 0.00
United Republic of Tanzania 4.62 4.00 3.50 2.93 3.09 3.68 3.12 –0.01 0.00d
Zambia 5.83 4.70 7.79 5.84 6.39 6.11 4.88 0.01 0.01d
Zimbabwe 9.87 12.26 9.72 6.38 4.99 8.64 6.44 –0.06 –0.04
Median 11.24 10.66 8.93 8.04 7.51 9.29 7.65 –0.04 –0.03
South and south-east Asia
Bangladesh 16.51 16.48 14.62 12.84 11.51 14.72 12.90 –0.05 –0.03
Cambodia 17.33 17.49 13.68 17.93 18.37 16.89 15.01 0.01d –0.01d
India 22.88 21.82 19.22 16.96 17.13 19.82 15.61 –0.05 –0.04
Nepal 12.26 14.51 11.91 9.36 7.53 11.46 9.69 –0.04 –0.03
Pakistan 18.97 12.47 9.16 12.03 7.88 12.56 9.21 –0.08 –0.05
Median 17.33 16.48 13.68 12.84 11.51 14.72 12.90 –0.05 –0.04
Eastern Mediterranean
Armenia 2.19 2.76 2.32 3.27 2.03 2.53 1.97 0.00d 0.01d
Egypt 3.33 3.41 3.20 2.89 2.82 3.17 2.52 –0.01d –0.01d
Kazakhstan 3.04 3.09 1.69 0.86 1.76 2.51 1.82 –0.01d 0.00d
Kyrgyzstan 3.21 3.43 4.11 3.16 1.06 3.28 3.44 –0.01d –0.01d
Morocco 14.22 9.34 9.87 9.19 10.52 10.74 9.31 –0.04 –0.05
Turkey 4.00 3.73 2.27 1.98 2.67 3.01 1.90 –0.01 0.00d
Uzbekistan 19.44 7.41 12.10 13.53 10.26 13.74 11.63 –0.08d –0.11
Median 3.33 3.43 3.20 3.16 2.67 3.17 2.52 –0.01 –0.01d
Latin America and the Caribbean
Bolivia 1.77 1.40 2.01 1.79 1.55 1.70 1.24 0.00d 0.00d
Brazil 4.41 2.48 2.24 1.41 2.64 2.75 2.34 –0.02 –0.01d
Colombia 1.74 1.69 1.68 1.27 1.12 1.54 1.29 –0.01 –0.01
Dominican Republic 3.16 1.90 2.77 1.88 1.44 2.15 1.70 –0.01 –0.02
Guatemala 2.76 3.86 4.21 1.10 2.71 2.91 2.52 –0.01d –0.01d
Haiti 8.09 5.40 5.91 4.05 5.52 5.81 4.61 –0.02d –0.01d
Nicaragua 3.86 2.23 2.78 0.87 1.66 2.37 2.07 –0.02 –0.01
Paraguay 0.73 0.56 0.47 0.67 0.39 0.56 0.33 0.00d 0.00d
Peru 2.16 1.02 1.03 0.72 0.71 1.15 0.94 –0.01 –0.01
Median 2.96 2.07 2.51 1.34 1.61 2.26 1.88 –0.01 –0.01d
Open in a new tab

a Based upon WHO Multicentre Growth Reference Study (MGRS).
b Derived using WHO and United States National Center for Health Statistics (NCHS) growth standards.
c Calculated as suggested by Erreygers.21
d Social inequality is not significant at the 10% level.

Table 2 and Table 3 also show average stunting and wasting rates, respectively, based on the new WHO child growth standards and the NCHS growth standards. For both indicators of malnutrition, the average rate was higher when the new WHO reference standards were used. However, socioeconomic inequality was fairly similar with the two different growth standards. Consequently, the following discussion relates mainly to the WHO child growth standards.

Fig. 1 plots the average level of stunting against the concentration index for socioeconomic inequality in stunting. For illustrative purposes, the negative of the concentration index is shown in this figure and Fig. 2, such that a high value on the y-axis indicates high socioeconomic inequality in favour of the rich. There was no clear association between average stunting and socioeconomic inequality in stunting (Spearman coefficient = 0.20, P = 0.17). If only socioeconomic inequality in the Latin America and Caribbean region was considered, there was an association between a high average level of stunting and high socioeconomic inequality in stunting.

Fig. 1.

Average stunting versus the negative of the concentration index

a Calculated as suggested by Erreygers.21

b Derived using WHO child growth standards.

Fig. 1

Open in a new tab

Fig. 2.

Average wasting versus the negative of the concentration index

a Calculated as suggested by Erreygers.21

b Derived using WHO child growth standards.

Fig. 2

Open in a new tab

Fig. 2 shows the same association for wasting and clearly illustrates that the socioeconomic inequality in wasting is much smaller than that in stunting. There was a negative association between average wasting and the concentration index for wasting (Spearman coefficient = –0.60, P < 0.001), which implies that countries with higher average wasting tend to have higher socioeconomic inequality. However, Fig. 2 also shows that the magnitude of the association was low, at best. The low values of the concentration index for socioeconomic inequality, combined with the finding that the relative variability in the average wasting level across countries (coefficient of variation = 0.68) was higher than that in the average stunting level (coefficient of variation = 0.35), suggest that one should not focus too much on the significance of the association between average wasting and socioeconomic inequality in wasting.

When the traditional concentration index (or the one suggested by Wagstaff)26 was used, different results were obtained for these associations. That is, there was a strong positive association between average stunting and socioeconomic inequality in stunting (Spearman coefficient = 0.78, P < 0.001), whereas the association between average wasting and socioeconomic inequality in wasting was not significant (Spearman coefficient = 0.14, P = 0.35). This confirms the importance of correcting for dependence on the mean.

Table 2 and Table 3 also show the distributions of stunting and wasting for different socioeconomic status quintiles. The pattern of the distribution can vary, and this is illustrated for three selected countries in Fig. 3.29 In Rwanda, socioeconomic inequality in stunting could be characterized as “mass deprivation” – stunting is highly prevalent within the majority of the population while a small privileged class is much better off. A second pattern, as seen in Ghana, could be described as “queuing” – average stunting is lower than in the previous pattern, but richer population groups are better off while the poor have to wait for a “trickle-down” effect. Third, socioeconomic inequality in stunting in Brazil took the form of “exclusion” where the prevalence of stunting is relatively low in the majority of the population but was much higher in a poor deprived minority.

Fig. 3.

Distribution of stunting across socioeconomic status quintiles for three countries

a Derived using WHO child growth standards.

Fig. 3

Open in a new tab

Discussion

This study illustrates that socioeconomic inequality in malnutrition is present throughout the developing world. The study findings show that the better-off suffer less from malnutrition and that the resultant inequality is much more pronounced for stunting than for wasting. This finding could have been expected as previous evidence has suggested that socioeconomic status has a smaller effect on the stochastic conditions that precipitate wasting (e.g. unforeseen environmental factors and disease) than on long-term malnourishment.9,15 Socioeconomic inequality in stunting was largest in the Latin America and Caribbean region, with Guatemala being an outlier, which is also in line with previous findings.11,15,30

Average wasting and stunting rates derived using the WHO child growth standards were larger than those derived using the NCHS reference population. This was also found by de Onis et al. for Bangladesh, the Dominican Republic and a pooled sample of North American and European children.31 However, estimates of socioeconomic inequality in both stunting and wasting were similar with the different growth standards, as were the associations between socioeconomic inequality and average stunting or wasting.

When studying the association between average malnutrition and socioeconomic inequality in malnutrition, the choice of the inequality index used does matter. With Erreygers’ index,21 there was no clear association between average stunting and socioeconomic inequality in stunting (though some evidence for a limited association with wasting was found), while use of the traditional concentration index (or the one suggested by Wagstaff)26 produced, instead, the opposite findings. It is worth noting that Wagstaff and Watanabe15 found evidence for an inverse relationship between being underweight and socioeconomic inequality on using the traditional concentration index. Applying Erreygers’ index to the data in their paper reverses this finding, which illustrates Erreygers’ point about the need to be careful when comparing concentration indices across countries with very different stunting levels.

Socioeconomic inequality in stunting occurred in different patterns, which could be described as mass deprivation, queuing and exclusion. The manner in which systems based on primary health care will develop will be different in these different contexts. In the case of exclusion, programmes targeted at specific population groups, namely the poorest, are urgently needed to achieve pro-equity outcomes while in other instances, such as mass deprivation, a broad strengthening of the whole system, either alone or combined with targeting, is required.29

In this respect, the distribution of malnutrition across socioeconomic groups, as shown in Table 2 and Table 3, can provide a useful tool for health policy-makers as it can easily be used to classify countries according to the above-mentioned patterns.

There are several limitations to this study. First, it has to be noted that data were only available for children aged 0–3 years instead of 0–5 years for six of the 47 countries (i.e. the Central African Republic, the Comoros, India, Kyrgyzstan, the Niger and Togo). Since anthropometric deficits accumulate over time, average malnutrition rates for these countries were underestimated compared with rates for other countries. However, as already discussed by Wagstaff and Watanabe, changes in the age limit do not systematically produce an upward or downward bias in socioeconomic inequality.15 Furthermore, the results were found to be robust when these countries were excluded.

Second, the use of an asset index to capture socioeconomic status has its shortcomings. Houweling et al. have shown that the choice of asset can influence the observed magnitude of the health inequality, but also conclude that, in the absence of reliable information on income or expenditure, the use of such an asset index is generally a good way of distinguishing between socioeconomic layers within a population (see also Wagstaff and Watanabe).32,33 With respect to the present study, it is important to note that a separate asset index was constructed for each country. It was, therefore, possible for the correlation between assets and socioeconomic status to vary between countries.

Third, the present study investigated only socioeconomic inequality in childhood malnutrition in the developing world and the extent to which inequality was related to the average malnutrition rate. Clearly, this is only a first step in a broader research agenda whose aim is to analyse the determinants of socioeconomic inequality in childhood malnutrition within and across developing countries. The next step should consist of combining literature findings on both socioeconomic and proximate determinants of malnutrition, such as feeding practices, health-care seeking behaviour and the mother’s nutritional status (e.g. Smith et al., Mosley and Chen, and Ruel et al.),17,34,35 with a decomposition approach, such as the one proposed by Wagstaff et al.10

Conclusion

The findings of this study have both methodological and policy implications. With regard to methodology, this paper is the first to study socioeconomic inequality in childhood malnutrition in the developing world using recently introduced WHO child growth standards. It was found that, although average malnutrition is higher when using this reference population, estimates of socioeconomic inequality are fairly similar to those derived using the NCHS reference population. In addition, the analysis demonstrated that, when studying the association between average malnutrition and the concentration index, it is important to take into account the dependence of this index on the mean value of the binary malnutrition indicator. When this was done, there was no clear relationship between average malnutrition and socioeconomic inequality.

The absence of a relationship between average malnutrition and socioeconomic inequality also has important implications for health policy. It suggests that there was no fundamental difference in socioeconomic inequality between countries with a low average level of malnutrition and those with a much higher average level. While it is not clear from this study whether this is the consequence of a deliberate policy focus on average malnutrition levels, it does indicate that policy-makers should be aware that a focus on reducing the average malnutrition level does not seem to lead to obvious generalized benefits. Nevertheless, the main goals and targets of large-scale development programmes such as the Millennium Development Goals continue to be couched in terms of improving population averages.36

The results of this study indicate that not only the degree of socioeconomic inequality in malnutrition but also its pattern should be of concern in setting health policies. To reduce malnutrition in, for example, a range of Latin American countries, policies should be targeted at the poor. In contrast, in many sub-Saharan African countries, there is substantial scope for progress by focusing simply on the general population, in addition to targeting the poor. ■

Acknowledgements

Many thanks to Guido Erreygers for stimulating discussions on this topic. Ellen Van de Poel acknowledges the University of Antwerp and the World Health Organization for support and funding. Niko Speybroeck is also affiliated with the Public Health School, Faculty of Medicine, Université Catholique de Louvain (Belgium). Tom Van Ourti is a Postdoctoral Fellow of the Netherlands Organisation for Scientific Research – Innovational Research Incentives Scheme – VENI and a member of the Tinbergen Institute (the Netherlands).

Footnotes

Competing interests: None declared.

References

  • 1.Bryce J, Boschi-Pinto C, Shibuya K, Black RE.and theWHO Child Health Epidemiology Reference Group. WHO estimates of the causes of death in children. Lancet 20053651147–52. 10.1016/S0140-6736(05)71877-8 [DOI] [PubMed] [Google Scholar]
  • 2.Murray CJL, Lopez AD. Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet. 1997;349:1436–42. doi: 10.1016/S0140-6736(96)07495-8. [DOI] [PubMed] [Google Scholar]
  • 3.de Onis M, Blössner M. The World Health Organization Global Database on Child Growth and Malnutrition: methodology and applications. Int J Epidemiol. 2003;32:518–26. doi: 10.1093/ije/dyg099. [DOI] [PubMed] [Google Scholar]
  • 4.de Onis M, Frongillo EA, Blössner M. Is malnutrition declining? An analysis of changes in levels of child malnutrition since 1980. Bull World Health Organ. 2000;78:1222–33. [PMC free article] [PubMed] [Google Scholar]
  • 5.Braveman P, Tarimo E. Social inequalities in health within countries: not only an issue for affluent nations. Soc Sci Med. 2002;54:1621–35. doi: 10.1016/S0277-9536(01)00331-8. [DOI] [PubMed] [Google Scholar]
  • 6.Wagstaff A. Socioeconomic inequalities in child mortality: comparisons across nine developing countries. Bull World Health Organ. 2000;78:19–29. [PMC free article] [PubMed] [Google Scholar]
  • 7.Gwatkin DR. Health inequalities and the health of the poor: What do we know? What can we do? Bull World Health Organ. 2000;78:3–18. [PMC free article] [PubMed] [Google Scholar]
  • 8.Gwatkin DR. Poverty and inequalities in health within developing countries: filling the information gap. In: Leon DA, Walt G, eds. Poverty, inequality, and health: an international perspective Oxford: Oxford University Press; 2000. [Google Scholar]
  • 9.Zere E, McIntyre D. Inequities in under-five child malnutrition in South Africa. Int J Equity Health. 2003;2:7. doi: 10.1186/1475-9276-2-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Wagstaff A, van Doorslaer E, Watanabe N. On decomposing the causes of health sector inequalities with an application to malnutrition inequalities in Vietnam. J Econom. 2003;112:207–23. doi: 10.1016/S0304-4076(02)00161-6. [DOI] [Google Scholar]
  • 11.Larrea C, Freire W. Desigualdad social y malnutrición infantil en cuatro países andinos. Rev Panam Salud Publica. 2002;11:356–64. doi: 10.1590/S1020-49892002000500010. [Social inequality and child malnutrition in four Andean countries] [DOI] [PubMed] [Google Scholar]
  • 12.Thang NM, Popkin BM. In an era of economic growth, is inequity holding back reductions in child malnutrition in Vietnam? Asia Pac J Clin Nutr. 2003;12:405–10. [PubMed] [Google Scholar]
  • 13.Fotso JC, Kuate-Defo B. Measuring socioeconomic status in health research in developing countries: should we be focusing on households, communities or both? Soc Indic Res. 2005;72:189–237. doi: 10.1007/s11205-004-5579-8. [DOI] [Google Scholar]
  • 14.Hong R. Effect of economic inequality on chronic childhood undernutrition in Ghana. Public Health Nutr. 2007;10:371–8. doi: 10.1017/S1368980007226035. [DOI] [PubMed] [Google Scholar]
  • 15.Wagstaff A, Watanabe N. Socioeconomic inequalities in child malnutrition in the developing world [Policy Research Working Paper 2434]. World Bank, 2000. [Google Scholar]
  • 16.Pradhan M, Sahn DE, Younger SD. Decomposing world health inequality. J Health Econ. 2003;22:271–93. doi: 10.1016/S0167-6296(02)00123-6. [DOI] [PubMed] [Google Scholar]
  • 17.Smith LC, Ruel MT, Ndiaye A. Why is child malnutrition lower in urban than in rural areas? Evidence from 36 developing countries. World Dev. 2005;33:1285–305. doi: 10.1016/j.worlddev.2005.03.002. [DOI] [Google Scholar]
  • 18.WHO Child Growth Standards. Length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age Geneva: WHO; 2006. [Google Scholar]
  • 19.Garza C, de Onis M. Rationale for developing a new international growth reference. Food Nutr Bull. 2004;25(Suppl. 1):S5–14. doi: 10.1177/15648265040251S102. [DOI] [PubMed] [Google Scholar]
  • 20.Physical status: the use and interpretation of anthropometry [Report of a WHO Expert Committee; Technical Report Series No. 854]. Geneva: WHO; 1995. [PubMed]
  • 21.Erreygers G. Correcting the concentration index. J Health Econ E-published ahead of print. [DOI] [PubMed]
  • 22.Anonymous Use and interpretation of anthropometric indicators on nutritional status. WHO Working Group. Bull World Health Organ. 1986;64:929–41. [PMC free article] [PubMed] [Google Scholar]
  • 23.Filmer D, Pritchett L. Estimating wealth effects without expenditure data - or tears: An application to educational enrollments in states of India. Demography. 2001;38:115–32. doi: 10.1353/dem.2001.0003. [DOI] [PubMed] [Google Scholar]
  • 24.Van de Poel E, O’Donnell O, van Doorslaer E. Are urban children really healthier? Evidence from 47 developing countries. Soc Sci Med. 2007;65:1986–2003. doi: 10.1016/j.socscimed.2007.06.032. [DOI] [PubMed] [Google Scholar]
  • 25.Wagstaff A, Paci P, van Doorslaer E. On the measurement of inequalities in health. Soc Sci Med. 1991;33:545–7. doi: 10.1016/0277-9536(91)90212-U. [DOI] [PubMed] [Google Scholar]
  • 26.Wagstaff A. The bounds of the concentration index when the variable of interest is binary, with an application to immunization inequality. Health Econ. 2005;14:429–32. doi: 10.1002/hec.953. [DOI] [PubMed] [Google Scholar]
  • 27.Wagstaff A, van Doorslaer E. Measuring and testing for inequity in the delivery of health care. J Hum Resour. 2000;35:716–33. doi: 10.2307/146369. [DOI] [Google Scholar]
  • 28.O’Donnell O, van Doorslaer E, Wagstaff A, Lindelow M. Analyzing health equity using household survey data: A guide to techniques and their implementation Washington: World Bank; 2007. [Google Scholar]
  • 29.World health report 2003: shaping the future Geneva: WHO; 2003.
  • 30.Larrea C, Montalvo P, Ricaurte AM. Child malnutrition, social development and health services in the Andean Region [Research Network Working Paper R-495]. Washington: Inter-American Development Bank; 2005. [Google Scholar]
  • 31.de Onis M, Onyango A, Borghi E, Garza C, Yang H.for theWHO Multicentre Growth Reference Study Group. Comparison of the World Health Organization child growth standards and the National Center for Health Statistics/WHO international growth reference: implications for child health programmes. Public Health Nutr 20069942–7. 10.1017/PHN20062005 [DOI] [PubMed] [Google Scholar]
  • 32.Houweling TAJ, Kunst AE, Mackenbach JP. Measuring health inequality among children in developing countries: does the choice of the indicator of economic status matter? Int J Equity Health. 2003;2:8. doi: 10.1186/1475-9276-2-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wagstaff A, Watanabe N. What difference does the choice of SES make in health inequality measurement? Health Econ. 2003;12:885–90. doi: 10.1002/hec.805. [DOI] [PubMed] [Google Scholar]
  • 34.Mosley WH, Chen LC. An analytical framework for the study of child survival in developing countries. Popul Dev Rev. 1984;10:25–45. doi: 10.2307/2807954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ruel MT, Levin CE, Armar-Klemesu M, Maxwell D, Morris SS. Good care practices can mitigate the negative effects of poverty and low maternal schooling on children’s nutritional status: Evidence from Accra. World Dev. 1999;27:1993–2009. doi: 10.1016/S0305-750X(99)00097-2. [DOI] [Google Scholar]
  • 36.United Nations Millennium Development Goals. Available from: http://www.un.org/millenniumgoals/ [accessed on 18 February 2008].

Articles from Bulletin of the World Health Organization are provided here courtesy of World Health Organization

RESOURCES