Menu
Menu
Menu
Menu
Background: In Ethiopia, there is a paucity of studies on inequality in caesarean section using methodologically rigorous and well-established approaches. In this study, we showed extent and the overtime dynamics of inequality in caesarean section in Ethiopia following rigorous methodologies. Methods: The data for analysis came from Ethiopia Demographic and Health Surveys (EDHS) conducted between 2000 and 2016. We used the World Health Organization’s (WHO) Health Equity Assessment Toolkit (HEAT) to analyze the data. Caesarean delivery was disaggregated by four equity stratifiers, namely education, wealth, residence and regions. Relative and absolute summary measures were calculated for each equity stratifier to capture inequality from different perspectives. 95% Uncertainty Interval was calculated around a point estimate to measure statistical significance. Results: We found large socioeconomic and area-based inequalities in use of caesarean section in all study surveys. The inequalities have occurred in favour of socioeconomically advantaged women and those living in urban areas and certain regions such as Addis Ababa. While area-related inequality had generally increased with time, socioeconomic inequality showed fluctuation. Adoption of different measures in the study for the inequality analysis has caused the emergence of mix of patterns in caesarean section inequality over time. Conclusions: In all the surveys, wealthy and more educated women, and those residing in urban areas had higher chance of obtaining caesarean delivery. Policy makers should work to ensure caesarean section that is in the accepted normal range. More emphasis should be drawn to subpopulation with under use of caesarean section while at the same time, discouraging unjustified use of it.
Maternal death in Ethiopia is one of the highest in the world [15] and signals substantial inequality in access to maternal health care services [16, 17]. However, the country has remarkably reduced the burden of maternal and child mortalities over the last several years [15, 18]. Ethiopia had a Maternal Mortality Ratio (MMR) of 401 in 2017; MMR decreased from 298 deaths per 100,000 live births in 2000, to 401 deaths per 100,000 live births in 2017 [15]. The averted deaths may be linked to a concomitant rise in coverage of maternal and child health services interventions during the same time period [19]. However, the gains are not enough; the current maternal health care service coverage is far from the level required to achieve the SDG of 70 deaths per 100, 000 live births [20]. In 2015, Ethiopia adopted an equity-laden five-year rolling Health Sector Transformational Plan (HSTP) [21] to help facilitate implementation of effective interventions and ultimately meet the 2030 SDG. This study was performed using cross sectional data obtained from the Ethiopia Demographic and Health Survey (EDHS) conducted in 2000, 2005, 2011 and 2016. All were nationally representative household surveys conducted by the Central Statistical Agency (CSA) of Ethiopia. The EDHS is the only source of nationally representative data on several health indicators including caesarean sections in Ethiopia. Two interim EDHSs were conducted in 2014 and 2019, and both were not included in the analysis because they are not available in the HEAT software. The DHS collects data on a wide range of health topics such as maternal health, child health, domestic violence, female genital mutilation, HIV/AIDS, maternal and child nutritional status, to mention just a few. Samples in the DHS are deemed to be representative nationally as well as regionally and for urban and rural residence. The number of women age 15 to 49 years included in the surveys were 15,683, 16,515, 14,070 and 15,367 in 2016, 2011, 2005 and 2000, respectively. The sample share of each of the waves was presented in the results section (Table 1) [14]. Cesarean section rate among non-pregnant women aged 15 to 49 years disaggregated by education, economic status, place of residence and region, the EDHS between 2000 and 2016 a STATcompiler. The DHS Program. [Internet]. [cited 2020 Jul 13]. Available from: https://www.statcompiler.com/en/ The detailed sampling methodology of the DHS has been described elsewhere [22]. Briefly, DHS follows a stratified two stage clustered sample design. Stratification is done based on the sub-national regions and place of residence. Following stratification, samples are drawn from each stratum through a two-step process. In the first stage, census enumeration areas (EA) are selected through a sampling technique known as Probability Proportion to Size (PPS), where the probability of selection of an EAs is contingent upon its size (measured through number of households), and the larger the EA is, the higher its chance of being in the sample. Information on the number of EA is basically obtained from most recent census data. In the second stage, 25–30 households are selected systematically from each stratum. The surveys covered all women aged 15 to 49 years who gave birth in the 5 years preceding the survey. For the CS variable, all women who gave birth two or 3 years preceding the surveys were included. The inequality variable measured in this study is birth delivered through caesarean section. It is measured as the proportion of all births by caesarean section in the two or 3 years prior to the surveys. Restricting our analysis to two or 3 years prior to the surveys period allowed us to present a more recent status of the CS rate and its disparity. We disaggregated the caesarean birth by the four equity stratifiers: economic status, education, place of residence, and the subnational regions. The economic status (or wealth) has five categories: poorest, poor, middle, rich and richest. Educational status of the woman was classified as no-education, primary, secondary of higher; place of residence as urban vs. rural, and the sub-national regions included the nine regions and two city administrations. The disaggregated caesarean delivery (reported as a percent) was presented for each of the four EDHS time periods. Point estimates were calculated and presented with corresponding 95% Uncertainty Intervals. The educational status and wealth have a natural ordering and are known as ordered equity stratifiers whereas place of residence and regions are non-ordered equity stratifiers. Whether an equity stratifier is ordered or not affects the choice of summary measures to be calculated [14]. We used the 2019 updated version of the WHO’s HEAT software [23] for analyzing the socioeconomic and area-based inequalities in the CS rate in Ethiopia between 2000 and 2016. The software description and access have been described in detail elsewhere [24]. In summary, the WHO released the software in 2016 using free and publicly available R programming language and R packages. The software allows assessment of within country health inequalities of more than 30 indicators for the reproductive, maternal, newborn, and child health (RMNCH). It also permits bench-marking inequality in one country with that of another country, allowing direct comparison of inequality in two or more countries at the same time [23, 24]. Since the software is publicly available, there is no ethical concern regarding access. Type of health indicator of interest (favorable vs. adverse) and the inherent properties of dimensions of inequality determine choices and interpretations of summary measures for this inequality study. Caesarean section was disaggregated by the commonly used dimensions of inequality discussed above. In addition, we adopted summary measures of different use and statistical properties. We employed a combination of absolute and relative inequality summary measures. These were Difference (D), Population Attributable Risk (PAR), Slope Index of Inequality (SII) and Relative Concentration Index (RCI); the first three are absolute inequality measures and the last is a relative measure of inequality. These measures were calculated for each of the four equity stratifiers; for the wealth and education dimensions of inequality, we calculated all of the four inequality measures, and for the region and place of residence, we calculated the D and PAR. The detailed methods of calculation, interpretation and all other detailed properties of the measures employed in the study have been described elsewhere in detail [23]. We offer a brief description of properties and interpretations of the measures. Whilst positive values of the measures are indicative of a disproportionately higher coverage of the service among the advantaged sub-groups (women who have completed secondary education or higher, are wealthy, are urban dwellers and live in big cities such as Addis Ababa are deemed advantaged); negative values indicate the disadvantaged nature of the service. When the absolute inequality measures are zero or the relative inequality measures are one, there is no inequality. The higher the absolute value, the greater the inequality is. D is a simple measure suitable for showing the absolute difference between two categories within a dimension of inequality (i.e. urban vs rural for residence) or between 2 or more categories (i.e. regions) based on an identified reference subgroup in the category. The other three (PAR, SII, RCI) are weighted complex measures of inequality that take into account sizes of subpopulations used in the calculation, thereby producing estimates reflective of the subpopulation size [14, 23]. Complex measures are the one that take account of sizes of subpopulation used in the calculation of a measure in question, thereby producing estimates reflective of size of the subpopulation [14, 23]. The computation of SII and RCI was restricted to education and wealth dimensions of inequality since they require an ordered equity stratifier. To declare that CS shows statistically significant disparities across the sub-groups of each equity stratifer, and to determine whether or not the inequality changed with time, we computed 95% Uncertainty Intervals (UI) around point estimates of each measure for each survey. For all inequality measures, the lower and upper bounds of the UI must not include zero to interpret that inequality exists. We assessed the trend of inequality for each summary measure by referring to the UIs for the different survey years; if the UIs did not overlap, inequality existed. The findings were presented per the recommendation of the Strengthening Reporting of Observational studies in Epidemiology (STROBE) reporting guidelines [25]. Ethics approval is not needed as the data is retrospective, anonymous and publicly available. Ethical procedures were the responsibility of the institutions that commissioned, funded, and managed the surveys. All DHS surveys are approved by Inner City Fund (ICF) International as well as an Institutional Review Board (IRB) in the respective country to ensure that the protocols are in compliance with the U.S. Department of Health and Human Services regulations for the protection of human subjects.
N/A
