Disparities in the use of antenatal care service in Ethiopia over a period of fifteen years

The Maternal Mortality Ratio (MMR) in Ethiopia has declined by only 22% in the last decade from 871 per 100 000 live births in the year 2000 [1] to 676 per 100 000 live births in 2011 [2]. In order to achieve the 5th Millennium Development Goal (MDG) – of reducing maternal mortality by 75% in the year 2015 from the 1990 baseline level- access to adequate and high quality maternal health services including basic antenatal care (ANC), intrapartum care and postnatal care is essential [3].

ANC provides an opportunity for skilled providers so that they may identify and treat or refer complications [3]. Moreover, it was reported that in sub-Saharan Africa, women having four ANC visits are 7.3 times more likely to give birth in a health facility than those with no antenatal visits [4]. In Bolivia, the odds of modern contraceptive use of women with at least one ANC visit during pregnancy were about 4.3 times than those without [5]. A combination of a 60% prevalence of contraceptive use, improved transport and birth in a health facility can reduce 75% of the maternal deaths [6].

However, studies have documented that the ANC services are not equally distributed across the different socioeconomic groups. According to Barros and colleagues, in 54 of the 75 countries investigated in the Countdown to 2015 study, the coverage of antenatal care with a skilled provider was 77.9% (95% CI: 71.4-93.2), the rate ratio of use among the richest women relative to the poorest being 1.9 (95% CI: 1.2-1.9) [7].

Differences in antenatal care use were attributed to constraints in the supply side related to physical access, user fees and quality of the services. In a systematic review by Say and Raine, the reorganization of the health care system in Morocco has introduced fees for maternal health services that may have limited access to poor women [8].

Constraints in demand are equally important in determining the use of maternal health services, and may operate at the individual, household and community level. At the individual level, age, income, education and knowledge about the characteristics of, and need for health services may limit women from seeking maternal health care. In addition, culturally, women may have no autonomy to seek care when husbands or in-laws are conceived as the decision makers on care seeking practices for women [9].

As a strategy for tackling the supply side constraints -starting from 1996/1997- Ethiopia has started an ongoing ambitious national program, the Health Sector Development Program (HSDP). Key objectives of this program are universal coverage of primary health care and reduction of maternal mortality from 871 to 600 per 100 000 live births [10]. It is important to address disparities if the country is to achieve these targets.

So far, few studies have examined the determinants of ANC use in Ethiopia. In rural south Ethiopia, ANC use by women was inversely associated with age. In that study women aged 25–34 years were 43% less likely to use ANC compared to younger women [11]. Another study demonstrated lack of association between autonomy level in Ethiopian women, reflected by women being the sole decision-makers during daily household purchases, and use of ANC care [12]. According to a nationally representative study conducted in the 2000, urban residence and secondary/ higher education were found to be independent predictors of at least 1 ANC visit with odds ratios (OR) of 4.1 (3.1-5.5) and 4.0 (2.7-5.9), respectively [13]. This study did not control for economic status and work autonomy.

These studies were somewhat limited in scope involving single communities, focused on only one aspect of the inequity (women’s age or autonomy), or were not based on recent data with the control of major potential confounders. Therefore, an additional study of the factors contributing to inequity in the use of ANC service in Ethiopia was indicated. Furthermore, trends in an inequity in the use of ANC need to be investigated in order to inform policy debate.

The aim of this study was to better understand the disparities in ANC use across the urban/rural divide, the socioeconomic groups, the educational strata and regions of Ethiopia and the corresponding trends in inequity over time. Ultimately, this information will be used to target national and regional programs to the disadvantaged and marginalized groups of the Ethiopian society.

This study was based on nationally representative repeated cross-sectional surveys conducted by the Measure Demographic and Health Surveys (DHS) Corporation of the United States of America (USA) and the Central Statistical Authority (CSA) of Ethiopia in 2000, 2005, and 2011. These community-based surveys encompassed all administrative regions of Ethiopia, and included both urban and rural areas [1, 2, 14].

The current analysis included women of reproductive age (15–49 years) who have had at least one live birth in the five years prior to the time of interview, and who participated in one of the three surveys.

A two-stage stratified-cluster sampling was used for all the three surveys. In both the 2000 and the 2005 Ethiopia DHS, the sampling was based on the Census Enumeration Areas (CEAs) of the 1994 Population and Housing Census of Ethiopia [15]. First, a sample of 540 (139 urban and 401 rural) CEAs was selected using systematic sampling from all regions. Then, 27 and 24–32 households per CEA were selected systematically in 2000 and 2005, respectively. This resulted in a total of 14 642 households and 15 716 eligible women in 2000, and 14 645 households and 14 717 eligible women in 2005. The 2011 survey was based on the 2007 National Population and Housing Census of Ethiopia [16]. This survey has sought to select 624 CEAs (187 urban and 437 rural) resulting in 18 720 households and 17 385 eligible women. Response rates for the first, second, and third surveys were 97.8% (n = 15 370), 95.6% (n = 14069), and 95% (n = 16 515), respectively.

In-person interviews were conducted by trained interviewers using structured questionnaires. As mentioned above, women were eligible for this study if they had at least 1 live birth in the five years preceding the survey. If a woman had more than 1 live birth in that period, she was asked about the most recent live birth (index pregnancy).

In this nationally-representative study, use of ANC services in Ethiopia has substantially increased since 2000. However, there is evidence of widening gaps by levels of education and economic status. Furthermore, the gap between Addis Ababa/Dire Dawa regions and Amhara region has increased during the study period, whereas gaps between Amhara region and most other regions have almost disappeared. Differences by urbanity remained wide with no clear trend.

Our finding of a gap in ANC use based on economic status is consistent with a study from Nigeria, which measured economic status by household assets and found that women from the very rich households were 6 times more likely to use ANC compared with women from the very poor households [20]. Similarly, a study from India has shown women from the highest economic status households to have a 7-fold increase in ANC visits compared with their counterparts from the lowest economic status households [21]. However, in Namibia the rate ratio of the rich to the poor was 1.06 [22].

In our study, in the year 2011, women with secondary or higher education had five times the odds of having ANC visits compared with women with no education, this is similar with findings from the Madhya Pradesh State of India, where the rates of having at least one ANC check-up of women with higher education (more than secondary) were about 5 times higher than among women with no education [23]. The variations across the levels of education could be due to the capacity of education to increase knowledge of health as well as cognitive ability, to locate needed services and demand quality health care once services are accessed, thus improving the use of health care services [24]. Education may also lead to increased autonomy and decision-making by women [25]. Nonetheless, economic status and work autonomy were controlled for in this analysis of the effect of education and therefore are unlikely to explain the whole association between education and the use of ANC.

Compared to rural women, women from urban areas had OR of 1.8-2.4 for ANC use, adjusted, among others, to wealth and education. Our findings of urban–rural disparities in ANC use are consistent with a study from Ecuador, where Ecuadorian women living in rural areas had an OR of 1.4 (95% CI: 1.0-2.0) for inadequate prenatal care compared with their urban counterparts [26]. However, a study from India reported an absence of independent association between places of residence and ANC use. In India, no urban–rural difference in ANC use was documented in some states of India such as Kerala, Tamil Nadu, Andhra Pradesh and Karnataka. This lack of association was attributed to the assignment of multipurpose health workers in rural areas [27].

The low rate of ANC use among rural women could be attributed to the uneven access to health care between urban and rural areas. As Ethiopia implements strategies to reduce maternal mortality, services are not adequately stretching to rural areas, partly due to lack of roads infrastructure, making distance a crucial factor that gauges the use of maternal health services. In a 2005 DHS report, two thirds of women aged 15–49 who have encountered problems in accessing healthcare cited concern having to take transport, distance to facility, no available provider, financial constraints and concern of the lack of female providers as the major barriers to accessing maternal health care [14]. The cost of the maternal health services could produce the financial constraints. According to Pearson and colleagues, 65% of the health centers they have surveyed in Ethiopia require an out-of-pocket payment for some aspect of care [28], which might be deterring the poor pregnant women from seeking care. Finally, health-seeking behavior of rural and poor women- which were not explored in this study- may have contributed to the urban/rural, economic and regional inequities.

The urban–rural divide had narrowed between 2000 and 2005. However, it was sustained afterwards. This could possibly be due to the initial dramatic expansion of health posts –which are located only in rural areas. In contrast to a 5 fold increase in the number of health posts between 2005 and 2009, there was a 38 fold rise in the number of health posts between 1997 and 2004 [29].

A large variation in the implementation of the HSDP and availability of maternal health services across regional states, and the historical concentration of health services around city states such as Addis Ababa and Dire Dawa may have contributed to the persisting differences across regions. The federal government of Ethiopia dictates that at least 4 health centers per 100 000 people are required in each region of the country. However, the number of health centers per 100 000 people is 0.43 in Somali, 1.56 in Afar, 1.64 in SNNPR and 2.2 in Amhara regions [30]. Even though the federal government has provided technical support to regions such as Afar and Somali, there was no specific budgetary support channeled to maternal health services [29]. This could be an additional mechanism for the wide regional variation in ANC use in Ethiopia, particularly between Addis Ababa/ Dire Dawa and other regions except Tigray.

Our study’s limitations include the inability to assess disparities in the use of 4 ANC visits –which is the new model recommended by World Health Organization [31], lack of assessment of cultural aspects such as acceptability, and attitudes of women and their partners towards ANC. Cultural preference might be a confounding factor, or otherwise could be an interacting variable, in the association between region, residence or economic status and ANC use. Women from different cultural backgrounds may have different beliefs and preferences regarding modern and traditional antenatal care. Lastly, quality of services, previous pregnancy outcomes and problems with on-going pregnancy which could potentially affect the use of ANC were not controlled for in our study.

Our study’s strengths include the nationally representative sample of women of reproductive age. Secondly, we have included community (region, religion and residence), individual (education, birth order and age) and household level (autonomy, economic status and marital status) determinants. In addition, we have controlled for both direct and indirect measures of autonomy, notably, wife beatings as one dimension and work status as a measure of another dimension of autonomy. These constructs are indicators of women’s position in the society as responsible members and their ability to spend their time for their own self-interest [32].

With regards to policy, the federal government should make effort to put more resources in rural areas and disadvantaged regions such as Somali and Afar. This should be compounded by maneuvers for reducing barriers to access -such as providing economic incentives for poor women or working on the quality of the services. The need for education of girls and women in Ethiopia as a strategy for improving ANC use cannot be overstated. Our study also highlights the need for further studies to explore additional factors which contribute to urban–rural and inter-regional disparities in the use of ANC services. We also recommend additional studies on the mechanism for the initial –between 2000 and 2005- narrowing of urban–rural disparities followed by a sustained gap afterwards, and the disparities in the use of 4 ANC visits in Ethiopia.

YEA is currently based at the Department of Health Services Management, Jimma University, Jimma, Ethiopia. His research interest mainly focuses on disparities in health among the different segments of the society that are unjust and unfair. He is also interested in studying the quality of health and medical care as it relates to effectiveness and efficiency. MRC is based at the Braun School of Public Health, Hadassah- Hebrew University, Jerusalem, Israel. She trains epidemiology and is interested in the determinants of maternal and neonatal mortality as well as foetal origins of adult disease.