1. Background
African cities have experienced tremendous population growth over the last few decades, and most of the future population growth in the region is expected to occur in urban areas [
1]. Unfortunately, this rapid pace of urbanization has been occurring amidst declining economies, leading to inability of local and national authorities to provide basic social services and employment opportunities to the growing urban population [
2]. Recent estimates show that urban population in sub-Saharan Africa (SSA) grew by almost 4.7% per year between 1980 and 2000 [
1], while per capita gross domestic product (GDP) dropped annually by nearly 0.8% [
3]. It is generally admitted that the impact of economic restructuring since the 1980s has been most severe on residents of major cities in SSA, following reduced public expenditure on municipal services, housing and infrastructure [
4]. Consequently, urban population explosion in developing countries and in SSA in particular, is accompanied by increasing urban poverty and malnutrition [
2,
5].
Newly assembled evidence from developing countries indicates that the locus of poverty and malnourishment is gradually shifting from rural to urban areas, as the number of urban poor and undernourished is increasing more quickly than the rural number [
6]. This trend is also illustrated by the narrowing urban-rural gap in child malnutrition in most countries of SSA [
7]. One of the distinct faces of urban poverty in SSA is the proliferation of overcrowded slums and shantytowns characterized by unhygienic environmental conditions (e.g. uncollected garbage, unsafe water, poor drainage and open sewers) which worsen the susceptibility of residents to various health problems [
2,
8]. As a result of such unhealthy conditions, rates of child malnutrition, morbidity and mortality are several times higher in slums and peri-urban areas than in more privileged urban neighborhoods, and even than in rural areas [
4,
9].
The evidence of large and even widening inequalities in health between the rich and the poor has stimulated international and national organizations to focus explicitly on the health and nutrition of the poor in the developing world [
10‐
12]. The focus on the poor is premised on the reality that the resulting poor health hinders human capital, thereby creating and perpetuating a vicious circle of poverty and poor health [
6,
13]. Thus, addressing the problems of inequalities in child health, both between countries and within countries, remains one of the greatest challenges, especially for policies and programs related to the Millennium Developments Goals (MDG) [
10]. The World Health Organization (WHO) corroborated the focus on improving the health of the most vulnerable and reducing inequalities between population subgroups and stated that
"the objective of good health is twofold: the best attainable average level, and the smallest feasible differences among individuals" [
14].
Against this background, the purpose of this paper is to contribute to the growing empirical literature on socioeconomic inequalities in health in developing countries, by examining differences across urban and rural areas in health inequalities. Specifically, the goals of this study are: (1) to document and compare the magnitude of inequities in child malnutrition across urban and rural areas; and (2) to investigate the extent to which socioeconomic inequalities
1 in urban areas are accounted for by the characteristics of communities, households and individuals. Given that urbanization has been one of the dominant underlying demographic processes in the past few decades not only in SSA, but also in the rest of the developing world, one of the key concerns is the extent of socioeconomic disparities in child health across urban and rural areas. Indeed, health-related resource allocation decisions generally rely on simple urban-rural comparisons, which mask the enormous disparities that are increasingly evidenced between socioeconomic subgroups in urban areas [
5].
The focus on malnutrition among children is predicated on the fact that undernutrition is one of the major public health concerns in developing countries, where it represents both a cause and a manifestation of poverty [
13,
15,
16]. The evidence of short and long-term consequences of nutritional deficiencies include increased risk of both morbidity from infectious diseases and mortality, impaired cognitive or delayed mental development and, subsequently, reduced learning abilities in school, and poor work capacity in adulthood [
17,
18]. Conversely, child undernutrition in developing countries is usually a consequence of poverty, with its attributes of low family income, poor education, poor environment and housing, and inadequate access to foods, safe water and health care services [
16,
19]. Investigating socioeconomic inequalities in child malnutrition within SSA is of special importance since the region is not on target to reach the MDGs. Recent data indicate that whereas malnutrition among preschoolers is substantially decreasing in Asia and Latin America and the Caribbean, it is on the rise in some countries of SSA, whilst in many others they remain disturbingly high or are declining only sluggishly [
17].
4. Discussion
This study has examined and documented differences across urban and rural areas in child health inequities. The first objective of the paper was to compare the scale of socioeconomic inequalities in child malnutrition across urban and rural areas. Our results show that in all countries and areas (urban or rural), children from the poorest households stand greater risk to be undernourished, than their counterparts in the most privileged households. Most studies that have used socioeconomic index [
21,
22,
25] or socioeconomic factors [
16,
18,
23] have reported similar results. More importantly, this study shows that while malnutrition is, on average, higher in rural compared to urban areas -a finding reported by other authors [
7,
43]- socioeconomic inequalities are, to a large extent, higher in cities than in rural areas. Many studies on socioeconomic inequalities in health have also shown evidence of higher heterogeneity of urban areas compared to rural settings, with the former harboring pockets of severe poverty and deprivation, and exhibiting substantial concentrations of ill-health among the poor [
5,
6,
9,
21].
Linking intra-urban disparities in Col. 4 of Table
3 to urban malnutrition in Table
2 shows that some countries like Mozambique, Nigeria and Uganda exhibit higher urban malnutrition rates and higher urban socioeconomic inequalities, whereas others like Ghana, Zimbabwe, Togo and Burkina Faso record lower values in both counts. Between these two extremes, Zambia, Chad, Madagascar, Tanzania, Côte d'Ivoire and Cameroon have lower values in one dimension and higher levels in the other. Results in Tanzania and Mozambique are worthy of attention. Despite its fastest urban population growth, Tanzania has a relatively low level of urban malnutrition, the largest urban-rural gap in malnutrition (see rural to urban odds ratio in Table
2), and a modest level of intra-urban inequalities in malnutrition. Like Tanzania, Mozambique witnessed faster urban population growth, coupled with increased per capita GDP. Yet, it has higher urban malnutrition, and more importantly, it records the largest intra-urban differences in child undernutrition. This finding indicates that the magnitude of within-urban inequities in child health is not merely a result of urban population growth, and suggests that well-designed policies can reduce these inequities even in countries facing urban explosion.
Another issue examined in this paper has been the magnitude of within-urban inequalities in child malnutrition across countries. Our results show large but varying levels of inequalities across countries, which are even larger than urban-rural differentials in malnutrition. Comparing within-urban differentials in child malnutrition to rural-urban differentials in malnutrition shown in Table
2 reveals that within-urban differentials are of higher magnitude compared to urban-rural differentials in all countries except Chad and Zambia, the only countries where the within-urban gap in stunting is not larger than the within-rural one. Indeed, rural to urban OR in the prevalence of child stunting vary from 1.2 in Madagascar to 3.0 in Tanzania with a median value of 1.6 in Uganda, whereas within-urban differentials in malnutrition range from 1.4 (Zambia) to 3.8 (Mozambique), for a median value of 2.3 (Burkina Faso), as indicated earlier.
This finding is in line with work of Menon
et al. [
5], which showed that intra-urban differentials in child stunting were larger than overall urban-rural differences in 8 out of 11 developing countries from SSA, Asia and Latin America. The fact that within-urban gaps in child health are larger than within-rural gaps, and even than overall urban-rural gaps, suggests that using global urban-rural prevalence to characterize child malnutrition may be misleading, since urban average could mask large differentials among socioeconomic groups in urban areas. These conclusions are in accordance with those of a number of studies which have demonstrated the existence of substantial concentrations of ill-health among the urban poor [
5,
9,
21]. They suggest that policies and programs geared at improving children's welfare should specifically include targeting the urban poor.
The third issue investigated in this work has been the extent to which within-urban differentials are explained by the characteristics of communities, households and individuals. Our data show that the influences of mother's and father's education, community SES, and bio-demographic variables are relatively modest in explaining inequities in child stunting among urban dwellers. This result corroborates findings from other studies which have demonstrated that household income is a key and independent determinant of food insecurity and malnutrition [
22,
44,
45]. The fact that adjusting for bio-demographic covariates produced an increase of urban inequities in most countries is quite surprising. Similar findings have been reported in other developing countries like Brazil where Sastry found that important differences in child mortality by place of residence were revealed by controlling for community characteristics [
36].
Limitations of the study
One of the problems in cross-country studies on urban/rural differentials is the classification of localities as urban or rural. Some countries classify in terms of administrative boundaries, others in terms of agglomerations. Other criteria used include population size, population density, or a combination of several of these criteria [
46]. Though this variety of urban/rural classifications undoubtedly weakens any cross-country comparisons, a uniform definition cannot capture the large variety of urban and rural situations across countries with such wide disparities of economic and social development as those used in this study. A second limitation of this analysis relates to our constructed community SES. Though the variable is worthy of interest given the growing body of research on the effects of neighborhood characteristics on health [
22,
37,
38], it should be noted that other community correlates likely to affect child health were not included in the analysis. These include variables that were not measured or not measurable such as food availability, agricultural and climate characteristics, air pollution, and epidemiologic data. The fact that community-level variance demonstrates statistical significance in all countries except Burkina Faso and Zimbabwe (not shown) is supportive of the possible effect of unobserved community factors.
5. Conclusion
This study has used standardized measures of SES defined at the household and community levels to document the scale of inequities in child malnutrition in SSA. It has shown that across countries in SSA, though socioeconomic inequalities in stunting do exist in both urban and rural areas, they are significantly larger in urban areas. Our results further show that intra-urban differences in child malnutrition are larger than overall urban-rural differentials in child malnutrition, and that they vary across countries, even among those with comparable levels of development. Finally, our results indicate that maternal and father's education, community SES and other measurable covariates at the mother and child levels only explain a slight part of the within-urban differences in child malnutrition.
Overall, the results of this piece of work suggest that specific policies geared at preferentially improving the health and nutrition of the urban poor should be implemented, so that while targeting the best attainable average level of health, reducing gaps between population groups is also on target [
14]. Haddad
et al. note that intra-urban differentials in health are not sufficiently highlighted [
6], and as Garrett & Ruel purposely point out, most programs to alleviate food insecurity and malnutrition are designed for rural areas, despite increasing evidence of declining living conditions in most cities of SSA [
44]. To successfully monitor the gaps between urban poor and non-poor, existing data collection programs, such as the DHS and other nationally representative surveys, should be re-designed to capture the changing patterns of the spatial distribution of population. Indeed, these programs usually exclude the slum areas since they are considered illegal settlements, and when they are included, the sample size is often too small to allow any reasonable slum specific estimates.
Notes
1In this paper the terms "socioeconomic inequalities" and "inequities" are used interchangeably. We do share the view that health inequality is a generic term used to designate differences and disparities in the health achievements of individuals and groups, whereas the term health inequities refers to inequalities that are unjust or unfair.
2HDI is a composite index based on three dimensions: health (longevity), education (literacy rate), and resource (standard of living). Countries are ranked in decreasing order of human development index (e.g. rank 1 corresponds to the highest human development level).
Declaration of competing interests
The author(s) declare that they have no competing interests.