Scholarly and policy interest in health disparities and in the relationship between inequality and health status has increased dramatically over the past few decades 123456, but most of it focused on the developed countries 7891011. Similarly, while research linking poor outcomes in child health to poverty and inequities was growing, much was still based in the developed countries 121314. However, the burden of child mortality and morbidity falls most heavily on the developing world 151617. As Black et al pointed out, "More than 10 million children die each year, most from preventable causes and almost all in poor countries" 16.

In the Arab region, several studies documented gender disparities in health, including female disadvantage in survival chances and access to health care for children and older adults 181920212223. However, empirical evidence regarding the link between socioeconomic (or social class) disadvantage and child health in the region remains tenuous.

Arab countries have made considerable progress in the past few decades in improving the health and wellbeing of their people. Infant mortality rates (IMR) for the region witnessed an impressive decline of about 69% from 1970 to the present, a sharper decrease than in most other developing regions 2124. However, despite such progress, or perhaps because of it, disparities within and across countries remained and may have even increased over time 2125.

Numerous socioeconomic factors play a role in determining child health, and the Arab region demonstrates considerable disparities in terms of socioeconomic and human development both within and across countries. Although intra-country socioeconomic disparities are more difficult to measure, in Tunisia – ranked by the UNDP as having a medium-level of human development – the richest 10% of the country's households absorbed 47.3% of all income/consumption, compared to only 2.3% by the poorest 10% 18.

Within the literature that exists, inter-country infant and child mortality disparities in the region have been specifically linked to wealth and socioeconomic status. Infant mortality for the region was found to be inversely correlated to female and male literacy rates, as was annual GNP per capita, the proportion of the population with safe drinking water, and adequate access to sanitation facilities 26.

Moreover, excess female child mortality has been prevalent in most if not all Arab countries, ranging from 10 to 30% 21. For countries with low levels of mortality, male children will normally have higher mortality rates than females 27. Thus, gender parity in child mortality indicates female disadvantage.

Health is more than survival and mortality, however, and should include consideration of factors that capture the deterioration of health, such as nutritional status and diseases, as well as measures that protect children's health, like immunization coverage. According to the WHO's Global Database on Child Growth and Malnutrition, child malnutrition in the region, as measured by stunting, has decreased on the whole since 1980 28. However, more than one in six children still suffered from malnutrition across the region, and one in 13 from severe malnutrition 21.

In the Arab region, intra-country socioeconomic disparities have also been shown to affect childhood nutritional status. Inequities like low household income, low maternal level of education, or residence in rural or squatter areas were shown in various studies to negatively impact child health indicators like under- or mal-nutrition, wasting or stunting in different Arab countries 2129303132.

This article provides a descriptive portrait of disparities in child health in 18 countries in the Arab region during the 1990s, using available tabulations and reliable micro data from national household surveys. Although descriptive, the paper attempts to answer three questions: (1) what was the extent of within-country disparities in child health in the region in the 1990s? (2) of the gender and socioeconomic disadvantages facing Arab children, which is greater?, and (3) is there an association between levels of health disparities and socioeconomic development of countries? An assessment of existing disparities in child health and the factors associated with them is an important undertaking if we are to arrive at appropriate remedial interventions designed to improve equity in health.

With the exception of child survival, gender disparities in child health demonstrated a female advantage. Although female deaths were well below those for males in nearly every country, the gap was not large enough when compared to a European standard. On the other hand, socioeconomic disparities in child survival and other child health indicators generally showed a large urban advantage, as well as an overall advantage for mothers with secondary education.

An urban advantage was evident in child survival, wasting, stunting and ARI, but not in immunization. Disparities by maternal level of education indicated an advantage for mothers with at least secondary education for child survival, fewer stunted and wasted children, and a greater immunization rate. There was no similar advantage to maternal education for ARI.

According to the literature, income remains worldwide as the most important determinant of child survival and wellbeing across and within countries 555657. A recent study on the socioeconomic determinants of infant mortality in 152 countries reported that GNI/capita, young female illiteracy, and income equality (as measured by the Gini index) were significant determinants of IMR; for middle-income countries, income equality was an independent predictor 5. Shawky 26 showed that in the Arab region a significant positive association existed between the level of income and an improvement in child health indicators.

Although the data did not show an absolute excess mortality of girls – rather seeming to indicate a male disadvantage – a comparison of gender disadvantage between the Arab countries and the European ones (used as a standard) showed that there was an excess female infant and under-five mortality in the region. Our findings correspond to those reached by Yount 22 in her study of sex differentials of child survival in the Middle East, with historical Northwest European life tables used as a standard. Our findings also correlated with those by Arnold 27. In his review of Demographic and Health Surveys from 26 developing countries (1986–89), Arnold found that the male infant mortality rate in 25 countries averaged to 19% higher than that of female infants; the rates for the Arab states, however, were much lower. Previous authors have linked a disparity in child survival indicators to inequality in seeking health services, disparity in food allocation and care and unequal treatment of boys and girls by health care providers 1923, all within a cultural context of son preference 20.

Female advantage in child nutritional status was documented. This can be explained by the biologically higher susceptibility of boys to diseases such as diarrhea and ARI, and their consequent failure to thrive. The male-female ratio in nutritional status rates should also be quantified in comparison to a standard, but such a standard was difficult to establish empirically. As cited by Arnold 27, in a 1990 UNICEF study of the wasting rates of boys and girls in 39 countries, the wasting of boys was found to be 1.3 times higher than that of girls. Our results showed that female advantage in nutrition in the Arab region was not as great. In particular, the Tunisian rate of wasting in girls (123/100) and Palestine's 116 stunted girls to 100 boys indicated a need to revise child nutritional policy.

Male disadvantage in ARI morbidity was also well documented. A male-female ratio of 1.25 has been reported in literature 58. The average male-female ratio for ARI in Arab region was lower, at 1.1. This difference might have resulted from boys receiving favoured attention to their illnesses. This favoured care was quantified by Ahmed 19 in a study establishing that the ratio for ARI was significantly in favour of boys as compared to girls. Given the established male disadvantage, the rates of Libya and Tunisia favouring boys (101/100 and 104/100 respectively), are of concern.

The absence of gender disparity in immunization coverage was reported in the literature 59. This achieved parity can be attributed to the generalization of immunization programs by most governments in Arab countries 59.

For disparity by place of residence, the levels in child mortality showed significantly higher rates in rural areas, as would be expected. The same trend was found in the nutritional status of children. Children living in rural environments showed a disadvantage in all indicators except for ARI. The urban penalty in ARI might have resulted from the higher rates of pollution in urban areas relative to rural places. This overall health disadvantage of rural populations in the Arab region was perhaps due to the lower socioeconomic status in rural settings that limit access to health care and adequate nutrition. However, it was difficult to separate these effects due to the lack of requisite micro data. There was no disparity by place of residence for immunization rates, again as a result of governmental policies in the region encouraging universal immunization coverage 60.

A significant disadvantage of illiterate mothers compared to mothers with at least secondary education was seen in wasting, stunting and vaccination rates. The disparities were higher than those found for gender and by place of residence. Although the benefits of literate mothers for children's health are well established, higher levels of education in mothers have also been linked to higher incomes 61. However, an education disadvantage was indicated in some Gulf States for wasting (Kuwait and Saudi Arabia) and for immunization (Bahrain, Oman and Saudi Arabia). A possible explanation of this finding could be the involvement of foreign domestic workers in care provision to the children of highly educated mothers 62.

Overall, this paper provided evidence of pervasive inequalities in child health in the Arab region. Although the findings point to an apparent female advantage in mortality rates and nutritional status during the 1990s in the region, these rates were well below what would be expected using a European standard rates indicating persisting female disadvantage. This persistent female disadvantage is perhaps due to persisting discriminatory practices against female children in the region. Rural residence and illiteracy were clearly markers of disadvantage in almost any measure of child health used in this study. It also appears that the gender gap was less important than that based on socioeconomic markers.

Associations between disparities with selected macro development indicators were all weak, and statistically non-significant. Thus, unlike overall levels of child survival and other health indicators, higher income or socioeconomic development did not seem to reduce common disparities in child health.

This evidence calls into question the global strategies adopted so far to improve child health in the Arab region. Unlike in the 1970s and 1980s, there was currently no 'magic bullet' to improve child survival 63, and mother's education may no longer impact infant survival in urban areas 64. Also, we know that economic growth or development may not be sufficient in generating health improvement of a population 65. The evidence presented here pertaining to child health, rather than merely his/her survival, has important implications for social and health policy at the regional and national levels. Policies which aim to improve access to public and private health services and to reduce gender and economic disparities may help improve children's health status and wellbeing 566667. Given current levels of infant mortality, greater attention should be directed at broader issues of child wellbeing, and to the struggle for more equity in health care provision.

The analysis reported in this paper has some limitations. First, there were some gaps in the data available to us from these survey reports. Estimates for the child health indicators used were not available for all 18 countries. Furthermore, mortality rates by maternal education were available for only a small number of countries, preventing us from conducting meaningful analysis of socioeconomic differentials in mortality. Second, the reliability of the estimates reported varied, depending on the survey design and competence of the field organizations responsible for data collections. Third, although the rates provided by these surveys were more accurate than the estimates and projections used by UN agencies, the surveys' raw data files were not available to us for undertaking more in-depth analysis. Thus, we were not able to examine disparities at the household level, incorporating relevant risk factors into the analysis. It is of particular interest, for example, to examine gender disparities in child health within socioeconomic strata, but such detailed tabulations were not available to us. Also, disaggregated child health indicators by socioeconomic variables such as income or household wealth were lacking. Finally, the conclusions regarding the links between disparities and socioeconomic development across countries were based on bivariate analysis owing to the small sample used here. More analytic research using multi-level statistical techniques incorporating both household and country level effects using micro data is needed, but must wait 'data democratization' by countries in the region.

The authors declare that they have no competing interests.

MK conceived of the study, supervised preliminary analysis and data compilation, conducted statistical analysis, and participated in writing the final draft. JD prepared the data, conducted preliminary analysis, and drafted sections of the first draft. SMN and RY conducted literature review and participated in analysis and drafting sections of the final manuscript. All authors reviewed and approved the final manuscript.