Data on poor-rich differences in under-5 mortality, full childhood immunization coverage, skilled delivery attendance and antenatal care for 43 low and middle-income countries were obtained from World Bank Country Reports [9 ]. The Country Reports are based on DHS data [10 ]. These are nationally representative surveys, for which usually between 5000–10000 women aged 15 – 49 years were interviewed. The data and indicators used have been described elsewhere in more detail [9 ]. We included those countries for which Country Reports were available at time of analysis.
Household wealth was the socio-economic characteristic used in this study. Wealth has been shown to be an important determinant of mortality and health care use. It is extensively used in the field of health inequalities research, especially in studies on low and middle-income countries. Wealth was measured using an index based on household ownership of assets. The assets were combined into a wealth index using Principal Components derived weights [9 ,11 (link)]. Despite its limitations [12 (link)], this index is fairly widely used as measure of economic status in developing countries [11 (link),13 ]. The total population in each of the countries was categorized accordingly into five, equally large, wealth layers.
First, scatter plots were used to assess the relationship between the overall level of the health-related outcomes and the magnitude of absolute and relative inequalities in these outcomes. The simplest inequality measures were used, i.e. the rate difference (RD) and the rate ratio (RR) between the poorest 40% and richest 40% population group. We calculated the R-square of the best fitting curve through the scatter plots.
Then, we examined to what extent the empirical patterns of the RR and RD could be clarified by mathematically-defined ceilings to the RR and RD. We calculated these ceilings using a hypothetical population of which 50% is poor and 50% is rich. For example, if overall immunisation coverage is 100%, the RR cannot exceed 1, and the RD cannot exceed 0. If overall immunisation coverage is 90%, the maximum value of the RR is 1.25 (i.e. 100% coverage among the rich and 80% among the poor) and is 20 for the RD. For outcomes that never reach 100%, like under-5 mortality, we made an adjustment to calculate realistic ceilings. We assumed a minimum under-5 mortality of 5 per 1000 live births and a maximum of 400/1000.
Household wealth was the socio-economic characteristic used in this study. Wealth has been shown to be an important determinant of mortality and health care use. It is extensively used in the field of health inequalities research, especially in studies on low and middle-income countries. Wealth was measured using an index based on household ownership of assets. The assets were combined into a wealth index using Principal Components derived weights [9 ,11 (link)]. Despite its limitations [12 (link)], this index is fairly widely used as measure of economic status in developing countries [11 (link),13 ]. The total population in each of the countries was categorized accordingly into five, equally large, wealth layers.
First, scatter plots were used to assess the relationship between the overall level of the health-related outcomes and the magnitude of absolute and relative inequalities in these outcomes. The simplest inequality measures were used, i.e. the rate difference (RD) and the rate ratio (RR) between the poorest 40% and richest 40% population group. We calculated the R-square of the best fitting curve through the scatter plots.
Then, we examined to what extent the empirical patterns of the RR and RD could be clarified by mathematically-defined ceilings to the RR and RD. We calculated these ceilings using a hypothetical population of which 50% is poor and 50% is rich. For example, if overall immunisation coverage is 100%, the RR cannot exceed 1, and the RD cannot exceed 0. If overall immunisation coverage is 90%, the maximum value of the RR is 1.25 (i.e. 100% coverage among the rich and 80% among the poor) and is 20 for the RD. For outcomes that never reach 100%, like under-5 mortality, we made an adjustment to calculate realistic ceilings. We assumed a minimum under-5 mortality of 5 per 1000 live births and a maximum of 400/1000.
Free full text: Click here
