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.

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.