Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.
Skip to main content
Download PDF

Level of and trends in women’s empowerment inequalities in antenatal care services in Ethiopia: further analysis of the Ethiopia demographic and health surveys, 2000-16

BMC Pregnancy and Childbirth volume 25, Article number: 102 (2025) Cite this article

Abstract

Background

Maternal health care services and women’s empowerment have received attention in the Sustainable Development Goals. Limited evidence exists on the extent of distribution of antenatal care services across the ladder of women’s empowerment in Ethiopia. In this study, we sought to shed light on whether and how such disparities changed over time.

Methods

Data for the study came from the 2000 and 2016 Ethiopia Demographic and Health Surveys. The outcome variables were three measures of antenatal care services: quality antenatal care, early antenatal care, and four or more antenatal care services. Women’s empowerment was measured through a newly developed index, SWPER Global. Specifically, we used two domains of the measure: attitude to violence and social independence. Disparities in antenatal care services were measured using the Erreygers concentration index, Relative Index of Inequality, Average marginal effect, and second difference of the average marginal effects. We decomposed the concentration index to study the contributions of different factors to the empowerment disparities in the services in 2000 and 2016 as well as to the over-time change in the disparities. The Oaxaca-type decomposition technique was applied to investigate social determinants’ role on the change in the disparities between 2000 and 2016. A generalized linear regression model was used for the analyses.

Results

According to the concentration index, women’s empowerment disparities in the utilization of antenatal care services existed in both surveys, where the services were disproportionately concentrated among women with better levels of empowerment. By the measure of average marginal effect, there were disparities favoring empowered women based mainly on the point estimates, except that the attitude toward violence disparity in 2016 occurred to the advantage of poorly empowered women. However, the confidence intervals suggest mixed findings. The concentration indices showed that disparities mostly increased in 2016 and the change was underpinned by the changes in the inequalities of various factors and sensitivities of antenatal care services with respect to these variables, such as wealth, maternal education, media exposure, place of residence, and women’s empowerment itself. Overall, the results of the second difference showed no large change in the disparities between 2000 and 2016. However, relative disparities decreased substantially during the same time.

Conclusions

While concentration index-based absolute inequalities increased, relative inequalities decreased, suggesting the importance of using both absolute and relative measures in a study. The decomposition analyses suggest that working on the equitable distribution of social determinants could improve empowerment disparities in antenatal care services.

Peer Review reports

Introduction

Within the broader context of health disparity, maternal and child health disparity is a particular issue. The rate of maternal deaths varies by, for example, racial belongingness and age of the mother, where black women and older women tend to experience more deaths [1,2,3]. Although socioeconomic injustices that have their roots in prejudice, racial oppression, and discrimination are the primary drivers of disparities in maternal health, the role of disparities in access to health care-along with disparities in other drivers like health insurance coverage-in contributing to poorer maternal health outcomes among people of color is well documented [2]. Even though inconsistent and inconclusive findings on the effectiveness of skilled ANC on maternal and birth outcomes were noted in literature [4], evidence shows that adequate ANC is linked with better maternal and perinatal outcomes, and lower levels of neonatal mortality [3, 5, 6]. Lack of access to adequate skilled ANC is linked with elevated odds of poor birth outcomes and neonatal mortalities [7, 8]. Besides this, during ANC sessions, pregnant women would receive a lot of health care services designed to prevent and treat infectious and non-communicable diseases [9].

Presently, however, the coverage of skilled ANC shows greater disparities within and between countries globally, with the sub-Saharan Africa (SSA) region showing the lowest coverage [9] as well as the highest disparities [10]. This shows that ANC coverage is far from being universal in most countries globally, especially in SSA settings, leaving room for ANC disparity to exist between women at various social hierarchies and positions in a population. Our survey of literature showed that disparities in access to skilled ANC are pervasive and occur according to many social determinants of health such as place of residence (urban vs. rural), women’s empowerment, occupation type, media exposure, age at birth, location (country or continent) and socioeconomic injustices [11,12,13,14,15,16]. Further, evidence unveils the stark differences in access to ANC between black and white women, and since higher rates of maternal mortality are concentrated among black women, it could be said that disparity in access to ANC could translate into disparity in maternal mortality though the unacceptably high mortality rate in black women can primarily be ascribed to inequalities in social determinant of health [2].

In Low-and middle-income countries, there is an increasingly large body of evidence on the association between women’s empowerment and use of maternal health care services including ANC [17,18,19,20,21,22,23,24], the associations generally ranging from preventive, null, and positive depending on how and where the studies were conducted. We noted that, in contrast to other socioeconomic variables, which are approximately identically measured across research and allow for comparison of findings, studies that evaluated the inequitable accessibility to ANC according to women’s empowerment used a variety of techniques to measure women’s empowerment. For instance, a study reported that 23 distinct variables of women’s empowerment were employed to evaluate their association with maternal health care services in the poor world [20], complicating both intra and inter-country comparability.

Not only is women’s empowerment being measured inconsistently and differently in the literature, no study evaluated the drivers of over time change of women empowerment disparities in ANC in Ethiopia. Using the recently released validated indicator of women’s empowerment, SWPER global, this study aimed at filling these gaps by following the analytic methods suitably developed for health care equity research. To our knowledge, one multi-country study applied this new index of empowerment in the study of ANC so far [11], but the article only provided descriptive information. Another study by Asim M applied the index to study its link with ANC in Pakistan [25]. A study by Ewerling F et al. used this index in the analysis of the Composite Coverage Index [26]. Our study used the new index to investigate women’s empowerment disparity in the utilization of ANC services and whether the disparity changed over time. Moreover, the study intends to examine whether the change was attributable to known and measured social determinants of health and health care. The Sustainable Development Goal (SDG) 17.18 calls for the disaggregation of available data by equity stratifiers found relevant in national contexts [27] to increase evidence for informed policy making. In this paper, we contribute to addressing the gap in the disaggregation of ANC services by women’s empowerment and go further to explain observed inequalities using advanced statistical methods. By doing so, we would add to the literature on how inequality in ANC be improved by addressing the disparity in empowerment status as well as other social variables.

Methods

Data source

For the study, data came from the Ethiopia Demographic and Health Surveys (EDHS) conducted in the 2000 and 2016. Where there are a few or no alternative options for obtaining data, such as in low- and middle-income countries like Ethiopia, the DHS program has shown to be of great use for data collection and research. The EDHS is a nationally representative household survey and collects data on many health issues such as maternal health, women’s empowerment, domestic violence, nutrition, and maternal and child mortality. The surveys focus mainly on women in the reproductive age range (15 to 49 years) and employ methodologies similar enough to allow comparability across time [28,29,30]. The final report of the respective surveys contains detailed information about the size of the samples and their selection procedure [29, 30]. Very briefly, samples were selected using a sampling procedure that is stratified two-stage cluster. After the nation was first stratified into urban and rural strata, samples were selected independently from each stratum in two stages. The probability proportional to size technique was used in the first stage to choose the enumeration area (EAs), which serves as the primary sampling unit. A fixed number of households were selected from each EA for the second stage through a systematic sampling procedure.

Measurements

Outcome variables

The study has three outcome variables, namely early first ANC, quality ANC, and at least four ANC services. All of the outcomes are skilled services provided by doctors, nurses, midwives, or health officers. We exclude HEW as it could not meet the criteria of skilled providers [31]. Early ANC is an ANC service started within the first three months of pregnancy. In the focused ANC model of the World Health Organization (WHO), completing ANC at least four times was required until the new ANC model, at least eight ANC visits, was introduced by the WHO in 2016 [32]. However, because the major DHSs in Ethiopia to date were conducted under the focused ANC approach, we chose four or more ANC rather than at least eight ANC. The ANC quality was created by using the ANC services pregnant women receive during their ANC attendance: whether or not blood and urine samples were taken and whether or not women had their blood pressure checked.

Information and services provided during ANC, such as information on potential pregnancy issues, nutrition advice, and assistance in creating a birth plan can serve as a proxy for quality ANC. However, in this study, we used only three of the services—blood pressure measurement, blood and urine testing because they were the ones that were consistently measured in both rounds of the surveys. Therefore, women who obtain all these three services are regarded as utilizing quality ANC. Services provided during ANC sessions are useful proxies for the quality of ANC services and elsewhere, similar studies have used different services to measure quality of ANC [11, 33]. All of the three outcome variables are binary, taking 1 if women receive the services and 0 otherwise.

Equity stratifier

Women’s empowerment serves as an equity stratifier in the study since we aim to investigate the empowerment of women-related disparities in the utilization of maternal health care services. We used an indicator that was recently created called SWPER Global, which serves as a useful overall indicator of women’s empowerment and was created for use in LMICs [34]. Developed using the Principal Component Analysis (PCA), the SWPER Global index is based on 14 different items available in the DHS and has three distinct dimensions: decision-making, social independence, and attitude toward violence. The domain of social independence measures the element of empowerment that enables women succeed in their goals, including education, information access, age at first-birth, and the educational and age differences between them and their husbands. The decision-making domain, on the other hand, assesses things like whether women participate in decisions about their health care, major household purchases, and visits to family or relatives, either alone or in conjunction with their husbands. Lastly, the attitude toward violence domain covers the question of whether it is ever appropriate to beat a wife. These sub-domains are each divided into low, medium, and high levels of empowerment and are ordinal-type variables amenable to studying inequality [34]. Due to the lack of data, however, we did not use the decision-making domain in our analysis.

Explanatory variables

In this study, we considered the following exposure variables: maternal age at pregnancy, place of residence, region, mother’s education, partner’s education, mother’s religion, media exposure, wealth index, mother’s occupation, and partner’s occupation. The exposure variables were selected to have the potential to operate as confounding variables in the association between women’s empowerment and ANC services and even if they do not, they would still function as independent predictors of the ANC services. Mothers are divided into two categories based on their age at pregnancy: 8 to 19 years old and 20 and older. Maternal education is classified into no education, primary, secondary/higher education. We lumped secondary and higher subgroups into one category due to sample size issues. Partner’s education is broken down into no education, primary, secondary, and higher education. Women’s religion is broken down into Orthodox, Protestant, Muslim, and traditional/other. Media exposure is broken down into no exposure, one media exposure, and two or more media exposures. The household wealth variable in the EDHS was created based on ownership of household items such as radio, bicycle, television, car, housing characteristics, water source, toilet facilities, and flooring/roofing materials of a household. Then using PCA, wealth is created out of the score produced and is expressed in five quintiles: poorest, poorer, middle, richer, and richest. We used the wealth index variable that comes pre-calculated in the dataset. Women’s and partners’ occupation status is broken down into yes (the mother or partner has occupation) and no otherwise. The variables are categorized according to the appropriateness of sample size in each subgroup of variable as well as similar classifications in previous studies.

Statistical analysis

To analyze the data, we employed various inequality analysis methods, namely, Erreygers Concentration Index (ECI), Average Marginal Effects (AME), second difference of AME, and Relative Index of Inequality (RII). We used a mix of absolute and relative as well as simple and complex measures of inequality following recommendations of the WHO [35]. The practice of mixing different inequality measures in a study has the potential to improve the credibility of findings [35]. The ECI and second differences of AME are absolute measures whereas RII is a relative measure. While the ECI and RII are complex measures, AME and second difference are simple measures because they make pairwise comparisons of health indicators between only two categories of a variable. The complex measures on the other hand take into account the whole population size represented in all categories of a categorical variable. That is, they use data from all subgroups to measure disparity. For example, for the wealth variable, AME and second differences compare ANC, say, between the poorest vs. the richest whereas ECI and RII produce a single summary measure that reflects the entire categories of wealth [35].

The concentration index (CI) is twice the area between the concentration curve and the line of equality. The CI has an inclusive range of -1 and 1. When the relevant health variable is predominately concentrated among the favored/better categories of a ranking variable, such as high empowerment in this study, the CI turns positive. Conversely, when the health indicator is predominately concentrated among the disadvantaged categories of the ranking variable, such as low women’s empowerment, the CI becomes negative. The CI becomes zero if there is no inequality. In addition to reflecting the entire range of the population studied, the CI combines many useful traits such as being sensitive to changes in the distribution of each subgroup of the ranking variable that would make it one of the most suitable measures of health inequality. The concentration index is defined as:

$$\:CI=\frac{2}{\mu\:}cov(h,r)$$
(1)

Where h denotes the health variable (ANC services), µ is the mean (proportion) of the health variable, r denotes the ranking variable (women’s empowerment) and cov denotes covariance between h and r. In the literature, various concentration indices are available to suit different health indicators to be measured. The standard CI is appropriate for health indicator variables measured on a ratio scale, and its usefulness for dichotomous indicators has been debated. For bounded health indicators such as ANC services, other variants of CI, such as Erreygers normalized concentration index has been suggested in the literature [36]. One of the problems facing researchers when applying the standard CI for bounded variables is that its value does not fall between − 1 and 1.

Consequently, this study used the ECI. It measures absolute inequality and has intriguing statistical qualities, such as meeting the mirror criterion, which states that inequality in attainment (getting ANC) is comparable to inequality in shortfalls (not receiving ANC), except for their signs. ECI is expressed as:

$$ECI = {{4\widehat y} \over {{\rm{ymax}} - {\rm{ymin}}}}CI$$
(2)

Where ymax - ymin is the range of the health variable, which is ‘one’ for binary variables, ŷ denotes the mean of the health variable (ANC care services). ECI is interpreted similarly to standard CI. A negative (positive) ECI indicates that ANC services are concentrated among women with lower (higher) levels of female empowerment. The conindex command was used to calculate the index [37].

We did decomposition analyses for the 2000 and 2016 surveys separately to investigate the influence of the predictors on the inequality in the use of ANC care services in the respective surveys. O’Donnell O et al. proposed a regression-based decomposition analysis technique and we used it in this study [38]. The ECI can be decomposed using the formula in Eq. 3.

$$ECI = 4\left[ {\beta j\overline x j*CIj + GCI{\rm{\varepsilon }}} \right]$$
(3)

Where \(\:\beta\:\)j is the partial effects of the set of determinants of ANC services (in dummy form), \({\overline x j}\) is the mean of the determinants, CIj is their concentration indices and GCIε is the generalized CI of the error term. Equation (3) produces elasticities (see below) and concentration indices of the variables in the model and their contributions to the empowerment disparity of ANC services. Concerning contributions, a variable widens (reduces) the disparity depending on whether it has positive (negative) sign.

We applied an Oaxaca-type decomposition approach to decompose the change in disparity in ANC services between the two survey points [38]. Since our response variables are binary, we used the GLM with the binomial family and logit link. The approach is used to calculate how much the shift in variables’ disparities and associations with the ANC services contribute to the over time change in the inequalities in ANC services. The decomposition of the change in ECI is formally computed as follows:

$$\eqalign{& \Delta ECI = {\rm{\Sigma j}}\,\eta jt\left( {CIjt - CIjt - 1} \right) \cr & + {\rm{\Sigma j}}\,CIjt - 1\left( {\eta jt - \eta jt - 1} \right) \cr & + \Delta \left( {{\rm{GCI\varepsilon t}}/{\rm{\mu t}}} \right) \cr}$$
(4)

Where t indicates survey period, \(\:\varDelta\:\) denotes first difference; \(\:\eta\:\) and CI refer respectively to elasticity and standard concentration index of the factors. The outputs of the decomposition of change, similar to the decomposition analyses for each survey as discussed above, include elasticity, concentration index, and contributions of the variables. Mathematically, elasticity for the binary ANC services is computed as:

$$\>(\beta \>{}^\prime {x^\prime }\>*\mu \>\_{}^\prime {x^\prime }\>)/\mu \>\_y$$
(5)

Where \(\:\beta\:`x{\prime\:}\) refers to regressors’ partial effect on ANC services, and \(\:\mu\:\_`x{\prime\:}\:\)and \(\:\mu\:\_y\:\) refer, respectively, to the mean of the regressors and ANC services. Elasticity measures the responsiveness of dependent variables (ANC services) to a 1% change in the independent variables [39]. Concentration index is a representation of the exposure variables’ concentration index. Depending on whether the value is positive or negative, the prevalence of the exposure variables is more common among the higher or lower categories of women’s empowerment, respectively. Lastly, the percentage contribution shows how much each variable in the model has contributed overall to the disparity in ANC services. A positive contribution for disparity that increased (decreased) over time shows that a variable contributed in the widening (reduction) of that disparity. Conversely, a negative contribution for disparity that increased (decreased) over time shows that a variable contributed in the reduction (widening) of the same disparity.

The RII is one of the indices used to measure social inequalities in health [40]. The goal of the RII is to measure the linear relationship between a health outcome of interest y, such as ANC services, and a ranking variable x, such women’s empowerment, in relative terms. It sums up the linear association across the entire scale of a ranking variable, such as women’s empowerment, using a single figure. A log-linear model of the form fβ (x) = y0 exp(βx), where y0 > 0 is used to define the RII. The empowerment gradient can be expressed as exp(β) when y = fβ (x). That is, the RII is defined to be exp(β), representing the size and direction of the linear association between x and y. The β is the least false parameter in the exp(β) function that produces the best estimation of the statistical relationship between the ranking variable, x, and the response variable, y. Our understanding of the RII as least false parameters necessitates the employment of a log-linear model for its computation no matter the form of the observed relationship between x and y. When the relationship between x and y is positive, the RII becomes bigger than one. When the association is negative, it takes values lower than one. When it takes the value 1, it indicates that there is no link between x and y. The detail accounts of the index is found elsewhere [40].

Women’s empowerment dimensions were transformed into ridit scores to make them suitable for assessing RII [41]. Mathematically, the ridit for a value x of a variable X (women’s empowerment in this study) is “equal to the probability that X < x plus half the probability that X = = x” [41] and we utilized the wridit Stata tool [42] to create the ridit scores. To investigate the change in the relative disparity over time, we assessed RII by interacting the transformed ridit scores of the women’s empowerment dimensions with a binary survey variable, indicating the 2000 and 2016 EDHSs in a pooled dataset. The interaction yields a single value that reflects the differential in empowerment in 2016 versus 2000. However, what we normally estimate is not RII, but ratio of the RII, henceforth referred to as rRII. Our understanding of the multiplicative interaction effect as rRII is guided by the literature, which reveals that in models incorporating interaction, what we measure is not relative risk and odds ratio, rather ratio of relative risk and ratio of odds ratio [40, 43, 44]. When the rRII is < 1, it suggests a reduction in the disparity in the second period compared to the first period. If it exceeds 1, the disparity widens over time. A modified Poisson regression model was fitted utilizing a GLM with Poisson family and log link for the computations of rRII. As the literature shows, compared to the binomial regression, the modified Poisson regression is shown to be less prone to issues of convergence, and it consistently and effectively predicts relative risks [45].

Several investigators and publications pay a lot of attention to the direction and statistical significance of effects, but frequently very little focus is placed on the substantive and useful implications of the findings [46]. Outputs of a regression model can be made more tangible and raise their interpretability by estimating predictive probabilities and marginal effects. For each survey, a GLM with logit link and binomial family was fitted that includes all the factors. Using the margins, mtable and mlincom Stata commands, AME was computed to get percentage point (pp) difference in predictive probabilities of ANC services between low and high subgroups of women’s empowerment for each survey. The same procedure was followed to compute second difference of AME, except that the GLM now includes, alongside the main effects, the interaction between women’s empowerment and survey. In nonlinear models, such as logit, interpreting the coefficient of an interaction term does not provide an accurate indication of whether differences exist between the two variables that interact. Instead, the second difference in AME is a more relevant measure [47]. Following this suggestion, second difference was computed in the study as the difference between AME in 2016 and that in 2000.

The GLM assumptions for both the Poisson and logit regressions hold, except that the assumptions related to study design were found to be violated. Specifically, the sampling process used to acquire the EDHS data is stratified and clustered, which violates the assumption of independent and identically distributed (IID) data that underpins traditional statistical analysis. As a result, the sampling strategies were accounted for the analyses to improve accuracy of our results. Thus, the three survey design components—stratification, clustering, and unequal probability of selection—were taken account of by using the Stata module svy. Our utilization of the survey design analyses additionally deals with the issue of variance overestimation that occurs when Poisson regression is used to the dichotomous outcomes. The analyses were done based on the pooled data drawn from the 2000 and 2016 waves of the EDHSs. To ensure that each survey counts equally in the analysis, the weight variable in the pooled dataset was rescaled. Because it is possible to make the mistake of using the same names for different strata and clusters in a pooled dataset drawn from more than one round of the DHS from a single nation, we created distinct stratum and cluster variables in the pooled dataset. The 2018 DHS statistics guide was used to handle missing and not-know replies [28]. When we created variables that had such replies, we used the missing and do not know responses in the denominator. Additionally, we recoded them into the variable’s “less advantageous” category (for example, into the category “no occupation” for the occupation variables). All analyses were done in Stata 16 software (Release 16. College Station, TX: StataCorp LLC).

Results

The background characteristics of the survey respondents are shown in Table 1. In all, 31,050 women of reproductive age from the two waves of the EDHSs were examined, with nearly identical percentages in both surveys. However, the actual sample size varies by the background characteristics of the respondents.

Table 1 Background characteristics of women of reproductive age, Ethiopia, 2000 and 2016 DHS
Full size table

Adolescent women made up between 13% and 16% of the sampled women in both surveys. In the 2016 EDHS sample, the percentage of urban residents grew from the 2000 EDHS by an average of nearly 4 percentage points (pp). The 16 years between the two surveys saw large changes in women’s educational attainment. Less than half of women in 2016 had no formal education, down from a 75.2% in 2000. On the other hand, there was a 19 and 8 points increase in the percentage of women with only primary and secondary or higher education, respectively.

The percentage of women with low levels of empowerment, in the attitude towards violence domain, decreased from 67.4% in 2000 to 46.9% in 2016, while the percentage of women with higher levels of empowerment climbed from 12.0 to 33.3%. Similar findings were found in the distribution of women’s social independence, which revealed that low levels of empowerment made up 70.4% of the sample in 2000 compared to 59.6% in 2016, while high levels of empowerment shifted from 5.7% in 2000 to 13.4% in 2016. See Table 1 for details.

Table 2 presents the percentage of ANC services in each subgroup of women’s empowerment in 2000 and 2016. The services were consistently higher in women with high empowerment status in both domains of empowerment, resulting in the empowerment gradient of the services. Also, the receipt of all the services increased over time for each subgroup of women’s empowerment. For example, ANC quality increased from 10.3 to 56.9%, from 19.8 to 59.2%, and from 33.1 to 67.0% in low, medium, and high subgroups, respectively, in the attitude to violence domains of empowerment of women.

Table 2 Coverage of ANC services by different domains of women’s empowerment, Ethiopia, 2000 and 2016 DHS
Full size table

Women’s empowerment disparities in the utilization of ANC services and their trends over time

The findings revealed that inequalities in ANC services exist across women’s empowerment dimensions, with differing magnitudes and time trends across the services as well as the dimensions of women’s empowerment considered.

The attitude toward violence differential in ANC quality had an ECI of 0.159 (95% CI: 0.114, 0.204) in 2000 and 0.096 (95% CI: 0.037, 0.155) in 2016, showing a reduction over time. The social independence gap in ANC quality, in contrast, increased in 2016 with an ECI of 0.191 (95% CI: 0.136, 0.246) from the 2000 level, which was 0.135 (95% CI: 0.086, 0.184). As of 2016, the ECI was 0.073 (95% CI: 0.044, 0.102), compared with 0.035 (95% CI: 0.017, 0.053) in 2000 for attitudes toward violence-based inequality in early ANC. The ECI for social independence was 0.039 (95% CI: 0.021, 0.057) in 2000 and 0.092 (95% CI: 0.061, 0.123) in 2016. The attitude towards violence disparity in four or more ANC was 0.065 (95% CI: 0.040, 0.09) in 2000 and 0.094 (95% CI: 0.061, 0.127) in 2016, and social independence disparity was 0.095 (95% CI: 0.071, 0.119) in 2000 and 2016, it was 0.142 (95% CI: 0.109, 0.175). The ECIs, except that of the attitude towards violence in ANC quality, showed that the disparities increased in the 2016 compared with that in 2000. Moreover, the ECIs were positive for all the services in both periods, showing that the services were disproportionately concentrated among women who had better empowerment.

The high women’s empowerment subgroup (compared with the low subgroup) in the attitude towards violence domain had an AME of 0.031 (95% CI: -0.019, 0.081) and − 0.022 (95% CI: -0.082, 0.038) on the probability of using ANC quality in 2000 and 2016, respectively. This means that women in the high empowerment group had, according to the point estimates, 3.1 pp higher and 2.2 lower chance of getting ANC quality in 2000 and 2016, respectively. Similarly, high women’s empowerment had an AME of 0.023 (95% CI: -0.046, 0.092) and 0.047 (95% CI: -0.040, 0.133), respectively, in 2000 and 2016 for the social independence domain, showing that women in the high empowerment subgroup had 2.3 pp and 4.7 pp higher chance of getting ANC quality than does the women in low empowerment group in 2000 and 2016. The high empowerment subgroup had AME of 0.014 (95% CI: -0.010, 0.038) in 2000 and − 0.004 (95% CI: -0.033, 0.024) in 2016 in the attitude towards violence and 0.013 (95% CI: -0.017, 0.042) in 2000 and 0.014 (95% CI: -0.030, 0.058) in 2016 for the social independence domain on early ANC. Finally, AME of 0.022 (95% CI: -0.014, 0.058) in 2000 and − 0.008 (95% CI: -0.042, 0.025) in 2016 in the attitude towards violence and 0.057 (95% CI: 0.008, 0.106) in 2000 and 0.024 (95% CI: -0.029, 0.077) in 2016 in the social independence domain were estimated to the four or more ANC.

The second differences in ANC quality between high and low empowerment in attitude towards violence and social independence were − 0.053 (95% CI: -0.131, 0.024), and 0.024 (95% CI: -0.076, 0.124), respectively. This means that the high vs. low empowerment disparity in the service decreased and increased by 5.3 pp and 2.4 pp in 2016 compared to 2000. Similarly, for the early ANC, the second differences were − 0.018 (95% CI: -0.055, 0.018) and 0.001 (95% CI: -0.045, 0.048), respectively, for attitude towards violence and social independence. This corresponds to the 1.8 pp going down and 0.1 pp going up in the inequality of the service. The second difference for attitude towards violence was − 0.03 (95% CI: -0.079, 0.019), and for social independence, it was − 0.033 (95% CI: -0.099, 0.032) for four or more ANC, suggesting 3 to 3.3 pp reduction in the disparity. Overall, the results of the second differences showed no large change between 2000 and 2016 in the disparities between the low and high subgroups of women’s empowerment.

On the other hand, according to the rRII’s results, there were fewer disparities in the utilization of ANC services in 2016 than in 2000. The rRII for the attitude towards violence disparity in ANC quality was 0.22 (95% CI: 0.13, 0.37) whereas, for the social independence inequality, it was 0.34 (95% CI: 0.20, 0.57). The rRII for the early ANC was 0.54 (95% CI: 0.30, 0.97) for attitude towards violence and 0.46 (95% CI: 0.26, 0.83) for social independence. Finally, the rRII for the four or more ANC were 0.45 (95% CI: 0.28, 0.73) and 0.33 (95% CI: 0.22 0.49) for the attitude towards violence and social independence-based disparities, respectively.

Decomposition of change in women’s empowerment inequalities in ANC services

The survey-specific decompositions of women’s empowerment inequalities in the utilization of ANC services are presented in Tables 3, 4 and 5. Women’s education (secondary or higher), wealth (mainly the richest subgroup), place of residence, both domains of women’s empowerment, media exposure, and region, such as mainly Addis Ababa, were among the factors that largely explained the empowerment disparity in ANC services in 2000 and 2016.

Table 3 Elasticities and concentration indices of factors and their contributions to women’s empowerment inequality in ANC quality, Ethiopia, 2000 and 2016 DHS
Full size table
Table 4 Elasticities and concentration indices of factors and their contributions to women’s empowerment inequality in early ANC, Ethiopia, 2000 and 2016 DHS
Full size table
Table 5 Elasticities and concentration indices of factors and their contributions to women’s empowerment inequality in at least four ANC, Ethiopia, 2000 and 2016 DHS
Full size table

Tables 6, 7 and 8 present the findings from the decomposition of the change in the disparities in the services. The reduction in the disparities of the services was linked with the change in elasticities and inequalities of the variables used in regression models. For example, the reduction in the attitude towards violence disparity in ANC quality was positively influenced by, among others, secondary or higher maternal education (567%), place of residence (245%), media exposure (245%), partner’s occupation (13%), and women’s occupation (8%). Other factors, such as age at pregnancy (-3%) had negative contributions. The other variables such as region, women’s empowerment, wealth, partner’s education and religion contributed either negatively or positively depending on whether contributions of categories of these variables had negative or positive signs. Overall, however, region (613%), attitude towards violence (165%), and religion (30%) had positive contributions, but partner’s education (-152%), wealth (-83%), and social independence (-53%) had negative contributions.

Table 6 Oaxaca-type decomposition of ECI for change in inequality in ANC quality, Ethiopia, 2000 and 2016 DHS
Full size table
Table 7 Oaxaca-type decomposition of ECI for change in inequality in early ANC, Ethiopia, 2000 and 2016 DHS
Full size table
Table 8 Oaxaca-type decomposition of ECI for change in inequality in at least four ANC, Ethiopia, 2000 and 2016 DHS
Full size table

Discussion

In this study, we investigated women’s empowerment-related disparities in the three measures of ANC services, and the contributions of different social determinants to the survey specific disparities as well as to the over time change of these disparities using the data drawn from the first and last waves of the EDHSs. The study showed that empowerment inequalities in ANC services occurred, often favoring highly empowered women in both surveys. Concerning the changes in the disparities between the surveys, the findings diverge depending on the measures used to examine the disparities. The second difference showed small but both increasing and decreasing patterns. The rRII, on the other hand, showed that gaps narrowed over time. Still, the pattern of disparities was different from the ECI measure, which revealed increased disparities, except that it showed a narrowing of attitude towards violence disparity in ANC quality. The study further unveiled that all the variables considered in the model contributed to the change in the women’s empowerment disparity in ANC services, with the contributions differing in magnitude and sign.

Globally, a large number of articles studied the disparity in ANC services according to different social determinants of health [11,12,13,14,15,16, 20]. However, the direct comparability of our findings would be complicated by the lack of studies that have examined the trends in disparity in ANC services in Ethiopia using the SWPER Global indicator of women’s empowerment. Although prior studies use different indicators of women’s empowerment and methods of analyses, our findings on the pro-empowered distribution of the ANC services are consistent with that in literature noted in developing world including Ethiopia [11, 13, 48]. However, we also found disparities though small in favor of women with low empowerment status in 2016 by the absolute measure of AME. The AME is a simple measure of inequality that compares ANC services in high empowerment with those in low empowerment, ignoring the ANC services in the medium empowerment status.

These divergent findings suggest that conclusions about inequalities in ANC and their trends can differ depending on whether one uses absolute or relative versus simple or complex measures. It is critical to note that the use of complex and simple as well as absolute and relative measures of health and health care inequalities in a single study is important to capture the whole story around a disparity. The WHO encourages equity researchers to use a mix of such measures of inequalities to convey findings from different perspectives [35].

Contribution of a variable to a social disparity in a health and or health care indicator is a result of the variable’s association with the indicator (elasticity) and its asymmetrical variations across levels of a ranking variable of interest (concentration index).

Consistent with findings reported in the literature, we showed the nexus between the use of ANC services and factors such as maternal as well as her partner’s education, wealth index, women’s empowerment, place of residence, region, and media exposure [22, 23, 48, 49]. The influence of these variables on the use of ANC services could follow some complex and interrelated pathways. Education is associated with health literacy [50, 51], suggesting a possible pathway for education to influence the receipt of maternal health care services. Further, cultural beliefs women hold are barriers to accessing early ANC services in certain African settings [52], and women’s education could potentially change such norms. Women’s empowerment could similarly affect women’s thinking and power as education is an important element of empowerment of women. Literature shows that women’s poverty is an important impediment to obtaining ANC services because poor women are not able to cover the costs associated with the services and transportation [52]. Moreover, rural women find it hard to get maternal health care services and this under- or non-utilization of services may be caused by a variety of obstacles, including bad road systems, trouble finding transportation, a great distance to go to a facility, and confusion regarding the facility’s regular hours of operation [53].

In our study, despite the disagreement of the findings across the measures, the decomposition of change showed that many variables played important roles in explaining the rise and reduction of women’s empowerment disparities in the ANC services. The contribution of education to the narrowing of the disparities in early ANC and four or more ANC may be related mainly to the greater fall in its elasticity in 2016. Studies shows that the gap in any ANC service use between women with no education and those who completed secondary or higher education decreased from nearly 51 pp in 2000 to about 41 pp in 2016 [29, 30]. The distribution of secondary or more education along the categories of empowerment was unfair in both the surveys, and the disparity concerning the social independence was even more unfair in 2016, as indicated by the rise of its CI, favoring women with high empowerment status. According to UNICEF (2022), female completion rates for lower and upper secondary schools are 22% and 14%, respectively [54]. Even though these statistics lag far behind SDG 4.1’s target of 100% [55], the literature shows that female education has improved in Ethiopia [29, 30]. The expansion of secondary education not only helps attain the universal coverage target of SDG, but it at the same time improves women’s empowerment since education is one of its components [34]. This would in turn translate into the reduction of education’s effect on women’s empowerment inequality in ANC (by decreasing its CI) and the rise of women’s empowerment itself.

Wealth, especially the richest subgroup, emerged as an important player in the growing of empowerment disparities in ANC services with time. Although it contributed to the rise of disparities in the ANC quality, wealth helped decrease disparities in the early and four or more ANC services between 2000 and 2016. Its role in the narrowing of the disparities could be related to the substantial fall of its elasticity though it became more unequally distributed in 2016 than in 2000. The reduction in the sensitivity of ANC services with respect to the richest wealth index may be linked to the increasingly better access to maternity care via, for example, the exemption policy on maternal health care [56]. However, health expenditure on maternal health care continues to create obstacles to women’s use of such services including ANC, with poorer women enduring the greater burden of out-of-pocket expenditure [57]. This may explain why the pro-empowered disparity in ANC services persisted into 2016, as these women tend to have a higher chance to be in the richest household. Literature showed that during the time the EDHSs were conducted, Ethiopia saw an increase in GDP from 6.1% in 2000 to 9.4% in 2016, with the highest of 13.6% recorded in 2004 [58]. Nevertheless, the differential distribution of the economic growth across women’s empowerment subpopulation may explain why wealth, especially the richest subgroup, in our study, concentrated more among women with high empowerment status, with the wealth inequality increasing between 2000 and 2016.

Women’s empowerment had a noticeable share for the change of empowerment disparities in the ANC services. The direction and strength of association of women’s empowerment has with ANC services varied by the domains of empowerment, types of ANC services, and by the survey period. It is important to note that the variations in what the domains of empowerment measure may explain the emergence of different results of the women’s empowerment inequalities in the services. The attitude towards violence domain, for example, reflects mainly issues related to violence against women such as wife beating whereas social independence measures such issues as education, mass media exposure, and age of women at first birth [34].

This mixed associations between empowerment of women and different maternal health care services is also reported in the literature [20]. Similarly, contribution of empowerment to inequalities varied by its domains and ANC services. For the inequality in early ANC, for example, the contributions of high subgroup of social independence domain were positive, suggesting its role in widening the inequality. This influence of the variable can be linked to the increment of its elasticities in 2016. That means women with high empowerment status became more capable of utilizing the services in 2016 than in 2000. On the other hand, the CIs showed a small improvement in the second survey period and this may reflect the improvement of empowerment of women over time in Ethiopia [59]. However, for the four or more ANC service, this domain of empowerment reduced the disparity, mainly due to the reduction of the elasticities, but the concentration index similarly decreased, again suggesting improvement of empowerment. The high subgroup of attitude towards violence domain, on the other hand, reduced the empowerment disparities in both the early ANC and four or more ANC. While the elasticity decreased greatly, the concentration index either reduced or remained unchanged, suggesting, respectively, progressively more utilization of the ANC services among women who had poor empowerment, and improvement of women’s empowerment.

The finding implies that even if empowerment was overall improved (see also Table 1), the works currently being done are not enough to empower women. Ethiopia’s 10-year development plan, for example, touches, among other issues, gender and social justice [60]. The plan however fell short of taking into account the context-specific nature of women’s empowerment and did not set objectives on how to empower women who are unequal at baseline. This “one-size-fits-all” approach to empowerment without recognizing that women are a heterogeneous population group could not just prevent them from being empowered, but may serve as an obstacle to accessing maternal health care services. Empowering women is a crucial strategy required to guarantee gender equality and is also connected to maternal health care services. Therefore, more has to be done to achieve women’s empowerment and gender equality as a goal in and of themselves [61] as well as to support the achievement of other SDGs, such as the target of universal access to reproductive health care services [62].

Moreover, the study showed that the distribution of women’s empowerment varied across other equity stratifiers like place of residence and region. Poorly empowered women, for example, live mainly in rural settings (have negative CIs). Similarly, the urban-rural disparity of empowerment of women was noted in the literature [59], suggesting that contextualized programs be needed to improve utilization of maternal health services in rural areas and hence narrow the empowerment disparities. A key area emerging from the current study is that inclusive economic growth and noticeable advancement in the education sector as well as national women’s machineries dedicated to women empowerment could drive the reduction of gaps in the utilization of ANC services between women at different levels of empowerment. The decomposition findings imply that the empowerment disparities in ANC services occurred due to the unequal distribution of multiple social factors, which calls for interventions that address the multiplicity of causes for the disparities.

The study has some strength. We applied advanced analysis method that helped us generate evidence that would help inform the development of policies needed to improve equity in maternal health care services. It also used different methods of inequality to assess the disparity from different perspectives. This would give policymakers the chance to see which methods lead to what findings and interventions would be developed by factoring in this issue. We used the new and validated indicator to measure women’s empowerment, SWPER Global, and comparability would be possible between studies. Also, our analyses were based on relatively large samples.

Our findings should, however, be interpreted in light of the quality of the data, contexts, and statistical methods used in the study, its design as well as broader contexts around this issue. We just analyzed associations and interpreted our findings as such and we caution users of the evidence to do the same, as establishment of causation requires more work, such as counterfactual analyses and qualitative investigation. Our analyses of ANC disparities were restricted to only married women as women’s empowerment requires that their partners’ variables be measured. Ethnicity was not used in the study due to sample size problems. Other potentially important variables not measured in the survey might have influenced the variable’s statistical relationship in the models. We did not analyze the decision-making domain of women’s empowerment as some variables used to create the index were not collected in the 2000 DHS.

Conclusions

In both surveys, women’s empowerment disparities existed in the utilization of ANC services to the disadvantage of poorly empowered women. Trends of disparities varied by the inequality measures, with absolute (ECI) and relative (rRII) measures, respectively, generally suggesting increasing and decreasing patterns between the 2000 and 2016. Several social determinants, like wealth, maternal education, place of residence, women’s empowerment, media exposure, and husband’s education were among the important contributors to the change in the absolute disparities. The study suggests that improvements in the equitable distribution of these social factors would help improve women’s empowerment inequalities in the services.

Data availability

The datasets generated and/or analyzed during the current study are available on the DHS website at http://www.dhsprogram.com.

Abbreviations

ANC:

Antenatal Care

CI:

Concentration Index

DHS:

Demographic and Health Survey

ECI:

Erreygers Concentration Index

EDHS:

Ethiopia Demographic and Health Survey

LMICs:

Low-and-Middle Income Countries

PCA:

Principal Component Analysis

PHC:

Population and Housing Census

PP:

Percentage Points

SSA:

Sub-Saharan Africa

WHO:

World Health Organization

References

  1. Hoyert DL. Maternal Mortality Rates in the United States, 2020 [Internet]. [cited 2022]. Available from: https://www.cdc.gov/nchs/data/hestat/maternal-mortality/2020/maternal-mortality-rates-2020.htm

  2. Hill L, Artiga S, Ranji U. Racial Disparities in Maternal and Infant Health: Current Status and Efforts to Address Them. KFF [Internet]. 2022 [cited 2023 Jan 15]; Available from: https://www.kff.org/racial-equity-and-health-policy/issue-brief/racial-disparities-in-maternal-and-infant-health-current-status-and-efforts-to-address-them/

  3. Kassebaum NJ, Barber RM, Bhutta ZA, Dandona L, Gething PW, Hay SI et al. Global, regional, and national levels of maternal mortality, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet [Internet]. 2016 Oct [cited 2023 Feb 18];388(10053):1775–812. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0140673616314702

  4. Hollowell J, Oakley L, Kurinczuk JJ, Brocklehurst P, Gray R. The effectiveness of antenatal care programmes to reduce infant mortality and preterm birth in socially disadvantaged and vulnerable women in high-income countries: a systematic review. BMC Pregnancy Childbirth [Internet]. 2011 Feb 11 [cited 2023 Jan 5];11(1):13. Available from: https://doi.org/10.1186/1471-2393-11-13

  5. Ghimire PR, Buck G, Jackson J, Woolley E, Bowman R, Fox L et al. Impact of Antenatal Care on Perinatal Outcomes in New South Wales, Australia: A Decade-Long Regional Perspective. Int J Environ Res Public Health [Internet]. 2023 Jan [cited 2023 Jan 5];20(2):977. Available from: https://www.mdpi.com/1660-4601/20/2/977

  6. Tekelab T, Chojenta C, Smith R, Loxton D. The impact of antenatal care on neonatal mortality in sub-Saharan Africa: A systematic review and meta-analysis. PLoS ONE [Internet]. 2019 Sep 13 [cited 2023 Jan 5];14(9):e0222566. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6743758/

  7. Arunda M, Emmelin A, Asamoah BO. Effectiveness of antenatal care services in reducing neonatal mortality in Kenya: analysis of national survey data. Glob Health Action. 2017;10(1):1328796.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Zhou H, Wang A, Huang X, Guo S, Yang Y, Martin K et al. Quality antenatal care protects against low birth weight in 42 poor counties of Western China. PLOS ONE [Internet]. 2019 Jan 16 [cited 2023 Jan 5];14(1):e0210393. Available from: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0210393

  9. UNICEF DATA. Antenatal care [Internet]. 2022 [cited 2023 Jan 5]. Available from: https://data.unicef.org/topic/maternal-health/antenatal-care/

  10. Barros AJ, Ronsmans C, Axelson H, Loaiza E, Bertoldi AD, França GV et al. Equity in maternal, newborn, and child health interventions in Countdown to 2015: a retrospective review of survey data from 54 countries. The Lancet [Internet]. 2012 Mar 31 [cited 2023 Jan 21];379(9822):1225–33. Available from: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(12)60113-5/fulltext

  11. Arroyave L, Saad GE, Victora CG, Barros AJD. Inequalities in antenatal care coverage and quality: an analysis from 63 low and middle-income countries using the ANCq content-qualified coverage indicator. Int J Equity Health [Internet]. 2021 Apr 17 [cited 2023 Jan 15];20(1):102. Available from: https://doi.org/10.1186/s12939-021-01440-3

  12. Arsenault C, Jordan K, Lee D, Dinsa G, Manzi F. Equity in antenatal care quality: an analysis of 91 national household surveys. Lancet Glob Health. 2018;6(11):1186–95.

    Article  Google Scholar 

  13. Joshi C, Torvaldsen S, Hodgson R, Hayen A. Factors associated with the use and quality of antenatal care in Nepal: a population-based study using the demographic and health survey data. BMC Pregnancy Childbirth [Internet]. 2014 Mar 3 [cited 2023 Jan 15];14(1):94. Available from: https://doi.org/10.1186/1471-2393-14-94

  14. Fagbamigbe AF, Idemudia ES. Assessment of quality of antenatal care services in Nigeria: evidence from a population-based survey. Reprod Health [Internet]. 2015 Sep 18 [cited 2023 Jan 15];12(1):88. Available from: https://doi.org/10.1186/s12978-015-0081-0

  15. Tekelab T, Chojenta C, Smith R, Loxton D. Factors affecting utilization of antenatal care in Ethiopia: a systematic review and meta-analysis. PLoS ONE. 2019;14:e0214848.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Tsegaye S, Yibeltal K, Zelealem H, Worku W, Demissie M, Worku A et al. The unfinished agenda and inequality gaps in antenatal care coverage in Ethiopia. BMC Pregnancy Childbirth [Internet]. 2022 Jan 29 [cited 2023 Jan 15];22(1):82. Available from: https://doi.org/10.1186/s12884-021-04326-y

  17. Ahmed S, Creanga AA, Gillespie DG, Tsui AO. Economic Status, Education and Empowerment: Implications for Maternal Health Service Utilization in Developing Countries. PLOS ONE [Internet]. 2010 Jun 23 [cited 2023 Jan 21];5(6):e11190. Available from: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0011190

  18. Chol C, Negin J, Agho KE, Cumming RG. Women’s autonomy and utilisation of maternal healthcare services in 31 Sub-Saharan African countries: results from the demographic and health surveys, 2010–2016. BMJ Open [Internet]. 2019 Mar 1 [cited 2023 Jan 21];9(3):e023128. Available from: https://bmjopen.bmj.com/content/9/3/e023128

  19. Sripad P, Warren CE, Hindin MJ, Karra M. Assessing the role of women’s autonomy and acceptability of intimate-partner violence in maternal health-care utilization in 63 low- and middle-income countries. Int J Epidemiol [Internet]. 2019 Oct 1 [cited 2023 Jan 21];48(5):1580–92. Available from: https://academic.oup.com/ije/article/48/5/1580/5309028

  20. Pratley P. Associations between quantitative measures of women’s empowerment and access to care and health status for mothers and their children: A systematic review of evidence from the developing world. Soc sci med [Internet]. 2016[cited 2023 Jan 16];169:199 – 131. Available from: https://doi.org/10.1016/j.socscimed.2016.08.001

  21. Akram N, Hamid A, Akram MI, ROLE OF WOMEN EMPOWERMENT IN, UTILIZATION OF MATERNAL HEALTHCARE SERVICES: EVIDENCE FROM PAKISTAN. Pak Econ Soc Rev[Internet].2019 [cited 2023 Jan 21]; 57(1): 93–116. Available from: http://pu.edu.pk/images/journal/pesr/PDF-FILES/5-v57_1_19.pdf

  22. Kareem YO, Morhason-Bello IO, OlaOlorun FM, Yaya S. Temporal relationship between Women’s empowerment and utilization of antenatal care services: lessons from four National Surveys in sub-Saharan Africa. BMC Pregnancy Childbirth [Internet]. 2021 Dec [cited 2023 Jan 21];21(1):1–14. Available from: https://bmcpregnancychildbirth.biomedcentral.com/articles/https://doi.org/10.1186/s12884-021-03679-8

  23. Tessema ZT, Teshale AB, Tesema GA, Tamirat KS. Determinants of completing recommended antenatal care utilization in sub-Saharan from 2006 to 2018: evidence from 36 countries using Demographic and Health Surveys. BMC Pregnancy Childbirth [Internet]. 2021 Mar 6 [cited 2023 Jan 21];21(1):192. Available from: https://doi.org/10.1186/s12884-021-03669-w

  24. Ousman SK, Mdala I, Thorsen VC, Sundby J, Magnus JH. Social Determinants of Antenatal Care Service Use in Ethiopia: Changes Over a 15-Year Span. Front Public Health [Internet]. 2019 [cited 2023 Jan 21];7. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6603173/

  25. Asim M, Hameed W, Saleem S. Do empowered women receive better quality antenatal care in Pakistan? An analysis of demographic and health survey data. PLOS ONE [Internet]. 2022 Jan 6 [cited 2023 Jan 21];17(1):e0262323. Available from: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0262323

  26. Ewerling F, Wehrmeister FC, Victora CG, Raj A, McDougal L, Barros AJ. Is women’s empowerment associated with coverage of RMNCH interventions in low- and middle-income countries? An analysis using a survey-based empowerment indicator, the SWPER. J Glob Health [Internet]. [cited 2023 Jan 21];11:04015. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7979155/

  27. United Nations Statistics Division. Indicator 17.18.3 - E-Handbook on SDG Indicators - UN Statistics Wiki [Internet]. [cited 2023 Jan 28]. Available from: https://unstats.un.org/wiki/display/SDGeHandbook/Indicator+17.18.3

  28. Croft TN, Aileen MJ, Marshall CK, Allen, Guide to DHS Statistics [Internet]., Rockville et al. Maryland, USA: ICF;2018[cited 2023 Jan 16]. Available from: https://dhsprogram.com/data/Guide-to-DHS-Statistics/index.htm#t=Guide_to_DHS_Statistics_DHS-7.htm

  29. Central Statistical Authority/Ethiopia, Macro ORC. Ethiopia Demographic and Health Survey 2000 [Internet]. Addis Ababa, Ethiopia: Central Statistical Authority/, Ethiopia, and ORC Macro.; 2001. Available from: http://dhsprogram.com/pubs/pdf/FR118/FR118.pdf

  30. Central Statistical Agency - CSA/, Ethiopia ICF. Ethiopia Demographic and Health Survey 2016 [Internet]. Addis Ababa, Ethiopia: CSA and ICF; 2017. Available from: http://dhsprogram.com/pubs/pdf/FR328/FR328.pdf

  31. World Health Organization. Births attended by skilled health personnel [Internet]. [cited 2023 Feb 21]. Available from: https://www.who.int/data/nutrition/nlis/info/births-attended-by-skilled-health-personnel

  32. World Health Organization. WHO recommendations on antenatal care for a positive pregnancy experience[Internet]. Geneva, Switzerland: WHO; 2016[cited 2023 Feb 20]. 154p. Available from: https://www.who.int/publications/i/item/9789241549912

  33. Islam MM, Masud MS. Determinants of frequency and contents of antenatal care visits in Bangladesh: Assessing the extent of compliance with the WHO recommendations. PLOS ONE[Internet]. 2018[cited 2022 Oct 19]. Available from: https://doi.org/10.1371/journal.pone.0204752

  34. Ewerling F, Raj A, Victora CG, Hellwig F, Coll CV, Barros AJ. SWPER Global: A survey-based women’s empowerment index expanded from Africa to all low- and middle-income countries. J Glob Health[Internet]. 2020[cited 2023 Jan 15]; 10(2). Available from: https://doi.org/10.7189/jogh.10.020434

  35. WHO. Handbook on health inequality monitoring with a special focus on low- and middle-income countries. Geneva: WHO; 2013.

    Google Scholar 

  36. Kjellsson G, Gerdtham UG. On correcting the concentration index for binary variables. J Health Econ [Internet]. 2013 May 1 [cited 2023 Jan 16];32(3):659–70. Available from: https://www.sciencedirect.com/science/article/pii/S0167629612001737

  37. O’Donnell O, O’Neill S, Van Ourti T, Walsh B, conindex. Estimation of concentration indices. Stata J [Internet]. 2016 [cited 2023 Jan 16];16(1):112–38. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4819995/

  38. Donnell OO, Doorslaer EV, Wagstaff A, Lindelow M. Analyzing Health Equity Using Household Survey Data. A Guide to Techniques and Their Implementation [Internet]. Washington: World Bank; 2008[cited 2023 Jan 16]. 234 p. Available from: http://hdl.handle.net/10986/6896

  39. Openstax. Introductory Business Statistics. Interpretation of Regression Coefficients: Elasticity and Logarithmic Transformation [Internet]. 2022[cited 2023 Jan 16]. Available from: https://openstax.org/books/introductory-business-statistics/pages/13-5-interpretation-of-regression-coefficients-elasticity-and-logarithmic-transformation

  40. Moreno-Betancur M, Latouche A, Menvielle G, Kunst AE, Rey G. Relative Index of Inequality and Slope Index of Inequality: A Structured Regression Framework for Estimation. Epidemiology [Internet]. July 2015 [cited 2024 Nov 1]; 26(4): 518–527. Available from: https://doi.org/10.1097/EDE.0000000000000311

  41. Brockett PL, Levene A. On a characterization of ridits. The Annals of Statistics [Internet].1977[cited 2023 Feb 23]; 5(6): 1245–1248.

  42. Newson R. WRIDIT: Stata module to generate weighted ridits [Internet]. 2021 July [cited 2024 Oct 30]. Available from: https://EconPapers.repec.org/RePEc:boc:bocode:s457427

  43. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ [Internet]. 2003 Jan 25 [cited 2024 Oct 30];326(7382):219. Available from: https://pmc.ncbi.nlm.nih.gov/articles/PMC1125071/

  44. Norton EC, Wang H, Ai C. Computing interaction effects and standard errors in logit and probit models. The Stata Journal [Internet]. 2004 [cited 2024 Oct 30];4(2): 154–167. Available from: https://journals.sagepub.com/doi/pdf/10.1177/1536867X0400400206

  45. Zou GA, Modified Poisson. Regression Approach to Prospective Studies with Binary Data. Am J Epidemiol [Internet]. 2004 Apr 1 [cited 2023 Feb 1];159(7):702–6. Available from: https://academic.oup.com/aje/article-lookup/doi/https://doi.org/10.1093/aje/kwh090

  46. Williams R. Using the Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects. Stata J [Internet]. 2012 Jun 1 [cited 2023 Jan 17];12(2):308–31. Available from: https://doi.org/10.1177/1536867X1201200209

  47. Mize T. Best Practices for Estimating, Interpreting, and Presenting Nonlinear Interaction Effects. Sociol Sci [Internet]. 2019 [cited 2023 Feb 15];6:81–117. Available from: https://www.sociologicalscience.com/articles-v6-4-81/

  48. Tiruneh FN, Chuang KY, Chuang YC. Women’s autonomy and maternal healthcare service utilization in Ethiopia. BMC Health Serv Res [Internet]. 2017 [cited 2023 Jan 21]. Available from: https://doi.org/10.1186/s12913-017-2670-9

  49. Yeo S, Bell M, Kim YR, Alaofè H. Afghan women’s empowerment and antenatal care utilization: a population-based cross-sectional study. BMC Pregnancy Childbirth [Internet]. 2022 Dec 27 [cited 2023 Jan 21];22(1):970. Available from: https://doi.org/10.1186/s12884-022-05328-0

  50. Harsch S, Jawid A, Jawid E, Saboga-Nunes L, Sørensen K, Sahrai D et al. Health Literacy and Health Behavior Among Women in Ghazni, Afghanistan. Front Public Health [Internet]. 2021 [cited 2023 Feb 21];9. Available from: https://www.frontiersin.org/articles/https://doi.org/10.3389/fpubh.2021.629334

  51. Gordon Singh S, Aiken J. The effect of health literacy level on health outcomes in patients with diabetes at a type v health centre in Western Jamaica. Int J Nurs Sci [Internet]. 2017 Jul 10 [cited 2023 Feb 21];4(3):266–70. Available from: https://www.sciencedirect.com/science/article/pii/S2352013216301429

  52. Chimatiro CS, Hajison P, Chipeta E, Muula AS. Understanding barriers preventing pregnant women from starting antenatal clinic in the first trimester of pregnancy in Ntcheu District-Malawi. Reprod Health [Internet]. 2018 Sep 21 [cited 2023 Feb 21];15(1):158. Available from: https://doi.org/10.1186/s12978-018-0605-5

  53. Ntoimo LFC, Okonofua FE, Igboin B, Ekwo C, Imongan W, Yaya S. Why rural women do not use primary health centres for pregnancy care: evidence from a qualitative study in Nigeria. BMC Pregnancy Childbirth [Internet]. 2019 Dec [cited 2023 Feb 21];19(1):277. Available from: https://bmcpregnancychildbirth.biomedcentral.com/articles/https://doi.org/10.1186/s12884-019-2433-1

  54. UNICEF. UNICEF Data. Monitoring the situation of children and women. Secondary Education and Enrollment Statistics [Internet]. UNICEF Global database on completion rate [Last updated 2022 June; cited 2023 Jan 21]. Available from: https://data.unicef.org/topic/education/secondary-education/

  55. United Nation. Sustainable Development Goals. Global indicator framework for the Sustainable Development Goals and targets of the 2030 Agenda for Sustainable Development [Internet]. [cited 2023 Jan 21]. Available from: https://unstats.un.org/sdgs/indicators/indicators-list/

  56. Federal Ministry of Health. Essential Health Services Package for Ethiopia. 2005[cited 2023 Jan 28]. Available from: http://dataverse.nipn.ephi.gov.et/bitstream/handle/123456789/1329/EssentialHealthServicesPackageforEthiopia%202005.pdf?sequence=1

  57. Akalu T, Guda A, Tamiru M, Haile Mariam D. Examining out of pocket payments for maternal health in rural Ethiopia: Paradox of free health care un-affordability. Ethiop. J. Health Dev [Internet]. 2012[cited 2023 Jan 29];26(1):251–257. Available from: https://www.ajol.info/index.php/ejhd/article/view/116112

  58. The World Bank. Data: Ethiopia [Internet]. [cited 2023 Jan 29]. Available from: https://data.worldbank.org/country/ethiopia

  59. Ministry of Women, Children and Youth and UNICEF Ethiopia. Gender Equality, Women’s Empowerment and Child Wellbeing in Ethiopia [Internet]. Addis Ababa: Ethiopia; 2020[cited 2023 Feb 1]. Available from: https://www.unicef.org/ethiopia/media/2811/file/Gender%20Equality,%20Women’s%20empowerment%20and%20child%20wellbeing%20in%20Ethiopia.pdf

  60. Federal Democratic of Ethiopia. Ten years development plan: a pathway to prosperity. 2021–2030. Addis Ababa: planning and development commission (ET) [Internet]. 2021[cited 2023 Feb 1]. Available from: https://europa.eu/capacity4dev/nexus_ethiopia/documents/ethiopia-ten-years-development-plan-2021-2030-planning-and-development-commission-federal

  61. United Nations. Transforming our world: the 2030 Agenda for Sustainable Development [Internet]. 2015[cited 2023 Feb 1]. A/RES/70/1. Available from: https://sdgs.un.org/sites/default/files/publications/21252030%20Agenda%20for%20Sustainable%20Development%20web.pdf

  62. Un women. Women and the Sustainable Development Goals (SDGs)[Internet]. [cited 2023 Feb 1]. Available from: https://www.unwomen.org/en/news/in-focus/women-and-the-sdgs

Download references

Acknowledgements

We appreciate the DHS program giving us access to the dataset.

Funding

The study was funded by Addis Ababa University. The funder has no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Author information

Authors and Affiliations

  1. Department of Reproductive, Family and Population Health, School of Public Health, Addis Ababa University, Addis Ababa, Ethiopia

    Gebretsadik Shibre & Wubegzier Mekonnen

  2. Department of Health Systems Management and Health Policy, School of Public Health, Addis Ababa University, Addis Ababa, Ethiopia

    Damen Haile Mariam

Authors
  1. Gebretsadik Shibre
    View author publications

    Search author on:PubMed Google Scholar

  2. Wubegzier Mekonnen
    View author publications

    Search author on:PubMed Google Scholar

  3. Damen Haile Mariam
    View author publications

    Search author on:PubMed Google Scholar

Contributions

GS conceptualized the design, analyzed and interpreted the data, wrote the draft manuscript. WM and DH supervised the design, analysis of data and the interpretations of findings. All authors reviewed and approved the final draft of the manuscript.

Corresponding author

Correspondence to Gebretsadik Shibre.

Ethics declarations

Ethics approval and consent to participate

We used data that is available in the public domain. The data from EDHS were collected after ethical approval was obtained. In addition, this study received ethical clearance from the institutional review board of Addis Ababa University.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shibre, G., Mekonnen, W. & Mariam, D.H. Level of and trends in women’s empowerment inequalities in antenatal care services in Ethiopia: further analysis of the Ethiopia demographic and health surveys, 2000-16. BMC Pregnancy Childbirth 25, 102 (2025). https://doi.org/10.1186/s12884-025-07223-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12884-025-07223-w

Keywords

BMC Pregnancy and Childbirth

ISSN: 1471-2393

Contact us