Background
Addressing inequities in maternal, newborn and child health (MNCH) is a key strategy to improve maternal, newborn and child health and survival, for the achievement of Millennium Development Goals (MDGs) 4 and 5 (reducing child mortality and improving maternal health, respectively) that are set for 2015 [
1], and for further improving MNCH beyond the MDGs. The Government of Ethiopia, recognizing the need to provide its people with equitable access to promotive, preventive and selected curative health services, launched the health extension program (HEP) in 2003. The HEP—a community—or
kebele-based health service delivery system—is the backbone of Ethiopia’s plan to achieve its priorities in the health sector, including the country’s health-related MDGs. (A
kebele is the smallest administrative unit of the country, with a population of about 5000.) The main strategies of the HEP include establishing a health post and training and deploying two government-salaried female health extension workers (HEWs) in every
kebele of the country. The HEP provides a package of services with 16 components in four major program areas: family health services, disease prevention and control, hygiene and environmental sanitation, and education and communication [
2‐
4]. Health centers, staffed with nurses, midwives and health officers, provide administrative, logistical, technical, and referral support to the HEWs. The health centers provide a wide range of mainly curative services and are being equipped to provide basic emergency obstetric and neonatal care. One health center provides support to five health posts and forms the primary health care unit, which is linked with a primary hospital to provide more specialized services including comprehensive emergency obstetric and newborn care [
5].
The HEP has achieved universal coverage by establishing at least one health post and deploying at least two HEWs in nearly all of the 15,000
kebeles in Ethiopia. The number of health centers supporting the health posts increased from about 800 in 2005 to more than 2000 in 2011 [
6]. At first, the HEWs spent 75 % of their time conducting household visits and community outreach activities, training families to adopt desirable health behaviors and practices and to serve as “model families” in their neighborhood, and organizing communities to participate in the expansion of HEP services. Families are taught hygiene, environmental sanitation, family planning, MNCH, disease prevention and control practices. Families that adopted 75 % of the healthy practices are said to ‘graduate’ as a ‘model family’ household [
7]. Community health volunteers referred as community health promoters (CHPs), who are from model family households and willing to volunteer have been supporting the HEWs in providing HEP services, with a density of one CHP for every 25 to 30 households [
2‐
4]. The HEWs and CHPs used a family health card (FHC), a booklet with pictorial messages, to promote focused MNCH care behaviors and practices.
Recently the government’s Ministry of Health implemented a new policy to increase the density of CHPs to one for every five households and to rename the CHPs as health development army members [
7]. With the initiation of integrated community case management of common childhood illnesses in 2011, the HEWs now spend about 50 % of their time at health posts [
8].
Since December 2008, the Bill & Melinda Gates Foundation has funded activities of the Last Ten Kilometers Program (L10K) implemented by JSI Research & Training Institute, Inc., to support the HEP to improve MNCH outcomes in 115 rural districts (
woredas), i.e., about three thousand
kebeles, of four regions of Ethiopia—Amhara, Oromia, Tigray, and Southern Nations, Nationalities and People’s—thus contributing toward the country’s achievement of MDGs 4 and 5. To do so, L10K project works to ensure that interactions between HEP front-line health workers—i.e., HEWs and CHPs (currently the health development army members)—and households to provide MNCH services will be more frequent, of higher quality, more cost-effective, and more equitable [
9].
The Ethiopia Demographic and Health Survey (EDHS) 2011 indicated that maternal and child health indicators have improved since the introduction of the HEP. Between 2005 and 2011, the contraceptive prevalence rate increased from 15 to 29 %, unmet need for family planning declined from 34 to 25 %, pre-natal care coverage increased from 28 to 44 %, deliveries assisted by skilled providers increased from six to 10 %, institutional deliveries increased from four to 10 %, births protected from neonatal tetanus increased from 32 to 48 %, and measles vaccination coverage increased from 29 to 56 %, while mortality in infancy and under age 5 declined from 77 and 123 deaths per 1000 live births, respectively, to 59 and 88 deaths per 1000 live births [
10]. Similarly, the L10K baseline and follow-up surveys (conducted in December 2008 and December 2010) conducted in the 115 L10K districts documented significant improvements in MNCH care behavior and practices [
9].
In order to provide universal primary health care, all services provided by the HEWs (and CHPs) are free of charge. The epicenter of the primary health care unit, i.e., the health centers, has user fees, but either MNCH services are provided free to all persons or user fees for these services are waived for the poor [
4]. It is critical to monitor whether the HEP has been able to establish an equitable health system to achieve universal primary health coverage. Culyer [
11] defines an equitable health system as one that “treats those with equal need equally and those in greater need … in proportion to that greater need”.
Traditionally, equity in global health has been measured by observing the differences in health care practices according to household wealth [
1], mainly because improving the health of the poor has been the top priority among international development agencies [
12‐
14]. However, according to the Culyer definition, equity in health can also have other dimensions such as age, education or residence (i.e., geographical location or distance from a health care provider). For instance, young women are at higher risk for adverse consequences of childbearing than women in their thirties, and children of young women are also at higher risk for morbidity and mortality and thus in greater need of MNCH services [
15‐
18]. Similarly, women with low levels of education also have greater need for MNCH services because they are at a higher risk of maternal morbidity and mortality, and the children of relatively uneducated mothers are likewise at higher risk of adverse health outcomes [
19,
20]. Furthermore, one HEP strategy for achieving universal coverage of primary health care has been to reduce the distance to service delivery points.
The community-based strategies of the HEP and L10K that influence universal and equitable accesses to priority MNCH services are listed in Table
1. Nevertheless, there are no studies to assess the adequacy of the HEP and L10K’s efforts to do so. Thus, this study uses data from the L10K surveys to assess the equity of use of MNCH services in L10K areas and whether there was any change in equity between the baseline and follow-up surveys. Equity of MNCH outcomes was examined along four dimensions: women’s age, education, wealth, and household distance from the nearest health facility.
Table 1
Community-based strategies implemented during the study contextual period that were aimed at establishing an equitable health system
- Establish one health post with two HEWs for every 5,000 populations - Free of charge - Health education during interaction with clients at health posts, communities and households - Train model families to adopt healthy behaviors and practices that influence their neighbor to do the same - Organize CHPs to promote HEP services - Community mobilization | - Provide refresher training to HEWs on maternal and newborn health - Train HEWs to organize CHPs to identify the target population for MNCH to promote and ensure HEP services, provide health messages to the target populations in hard to reach areas - Conduct review meetings with the HEWs at the woreda-level to exchange best practices - Provide supportive supervision visits to the HEWs to reinforce their skills and address performance gaps |
Discussion
This study examines the equity of four HEP outreach activity and 15 MNCH care practice indicators in the L10K areas according to women’s age, education, wealth, and distance from the nearest health facility, based on surveys conducted in December 2008 and December 2010, as well as changes in the inequities between the two survey periods. Inequities in the indicators of interest as of December 2010 were mainly due to age and education, followed by wealth and then distance from the nearest health facility. Although there were not many changes in the inequities in HEP outreach activity coverage and MNCH care practices between December 2008 and December 2010, the wealth inequity improved for three of the 19 indicators while it deteriorated for two of the indicators; the education inequity improved for two indicators and deteriorated for three indicators; the distance inequity improved for two indicators and declined for two indicators; while the age inequity deteriorated for six indicators.
L10K used the 30 by seven method mainly to minimize the cost of the survey field cost. However, the survey method is criticized because the interviewers may avoid hard-to-reach areas and non-responders are not revisited [
21]. The hard-to-reach population and the non-responders are likely to have comparatively low MNCH care behavior and practices; as such, it is likely this study under-estimated inequity. Nevertheless, the observed changes in inequities reported here are most likely unbiased because the sampling method was consistent between the surveys; as such, the survey method bias was held constant when the changes in equity were assessed. Although the 30 by seven sampling method used non-probability sampling to select households for interviewing the target respondents raising the question regarding the accuracy of estimating the 95 % confidence intervals of the point estimates [
30], a simulation exercise demonstrated that the accuracy of the variance estimates from such surveys are likely within the 95 % confidence limits [
31].
The study was an adequacy design; as such, the observed changes in inequities could not be differentiated from secular trend from program influence (i.e., L10K or HEP or both). The exploratory analysis of this paper tested a large number of hypotheses; as such, some of the statistically significant findings may be spurious.
Two methods for measuring equity (i.e., Wald’s statistics and concentration index) were chosen for this study from the various methods available, which are described elsewhere [
1,
26]. The concentration index value indicated whether the inequity was bottom or top; however, for Wald’s statistics the indicators according to the equity factor needed to be eye-balled to assess whether the inequity detected as statistically significant was top or bottom inequity. Only about half of the cross-sectional measures of statistically significant inequity during the follow-up survey corroborated between the two methods; while only six of the 20 statistically significant changes in equity measures observed between the baseline and follow-up surveys corroborated between the two methods. Since the different methods of measuring equity entail different sets of statistical assumptions, it would be advisable to use more than one method to measure equity and their changes over time.
The inequity in household visits by HEWs according to distance from the nearest health facility is of concern. However, it was reassuring to note that household visits by CHPs, model family households, and the possession of a family health card were not associated with distance from the nearest health facility, thus suggesting that the CHPs and model families are reaching the population in areas where HEW visits are relatively infrequent. It appears that maternal and newborn health care messages can reach areas where HEW visits are less frequent, as there were no inequities according to distance for 10 of the 15 maternal and newborn health inequity indicators analyzed in the follow-up survey. This finding is reinforced by the lack of inequity in possession of a family health card based on distance from a health facility.
Although there have been only two positive shifts in inequities in the indicators of interest according to distance from a health facility, it is encouraging to note that the proportion of the rural population who live more than one hour away from a health facility has been declining, from 22 % in December 2008 to just 9 % in December 2010. Thus, while those furthest from a health facility are still deprived of some aspects of MNCH care, they now represent a smaller fraction of the population.
Inequities in MNCH indicators according to women’s age existed for all four indicators of exposure to the HEP front-line workers. The L10K project supports the HEP in providing MNCH services. The expected pathway of the impact of L10K on MNCH behavior and practices is through the interactions of the HEP front-line workers with households. This equity analysis suggests that, to the extent that the impact of L10K occurs through HEW outreach activity, that impact would likely be inequitable according to women’s age, education and distance from a health facility, though not according to wealth. To the extent that L10K impacts MNCH behavior and practices through household visits by CHPs, then those impacts would likely be inequitable according to women’s age and education, but not by wealth or distance from a health facility.
This study has also shown significant improvements in the coverage of all the MNCH indicators other than that for treatment seeking behavior for ARI and the practice of ORT for managing diarrhea cases. It is understandable that there were no improvements in care seeking behavior for ARI because the provision for the service was not the part of HEP when the follow-up survey of this study was conducted. However, providing ORT had been within the HEP package of services. The HEP expects that there will be an improvement in the household care seeking behavior and practices for ARI and diarrhea case management following the expansion of the integrated community case management of common childhood illnesses in early 2011.
The existing strategies and policies of the HEP and L10K are less than optimum for minimizing inequities in MNCH services, whether according to women’s age, education, distance from a health facility, or wealth. The promotion of equitable MNCH services is essential if Ethiopia is to reach its MDG-related maternal and child health targets. Women age 15 to 19 and those who live more than an hour away from a health facility represent a small fraction of the target population for MNCH services (about 7 % and 9 %, respectively, during the follow-up survey). At the same time, the average distance from a health facility has been declining over time. Therefore, achieving equity of MNCH services by age and distance from a health facility would have only a modest (though still important) short-term impact on achieving the country’s overall health goals. For example, achieving equity in contraceptive use according to women’s age in L10K areas (i.e., enabling women age 15 to 19 to achieve the contraceptive prevalence rate exhibited by women age 20 to 34, shown in Additional file
1: Table S1) would mean an overall increase in contraceptive use of only 2 %, from 37 % (observed during the follow-up survey) to 39 %.
By contrast, the proportion of the women in rural L10K areas who have no education was substantial (78 %) in the follow-up survey. As such, addressing MNCH inequities according to women’s education will have greater impact and will significantly contribute toward achieving the country’s MDG targets related to maternal and child health. For example, achieving equity in MNCH indicators according to women’s education would mean an increase in the contraceptive prevalence rate from 37 % (during the follow-up survey) to 54 %. Similarly, institutional deliveries would increase from 11 % to 35 %, and deliveries assisted by health professionals would increase from 16 % to 41 %, among others.
The inequities in MNCH indicators according to education were most likely a combination of program uptake issues and differential targeting for MNCH services by the HEWs. Better-educated women may be more likely to proactively seek out and accept MNCH services provided by the HEP; as such, less educated women are lagging behind. Therefore, efforts to address inequity in the health sector should be complemented by other efforts in the broader social sector, where the introduction of strategies to reduce illiteracy and improve the population’s education will eventually eliminate education as a major source of inequity.
Assessing MNCH service equity due to religion and ethnicity was beyond the scope of this study. Nevertheless, there may be other health inequities related to these and other factors. Measuring such inequities is important for monitoring the effectiveness of existing policies for universal social services in Ethiopia. Accordingly, the equity of MNCH services by religion, ethnicity, region and place of residence should also be monitored using the EDHS.
One of the major assumptions for assessing equity of the HEP in providing MNCH services was that certain segments of the population (for instance, women less than 20 years old, women with no education, or women in the poorest health tercile) have greater health needs or less access to health services than others. However, this assumption was not tested. Future studies on equity of HEP services should test this assumption or assess equity according to health need.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AMK conceived and designed the study, carried out the statistical analysis, and drafted the paper; AT, AAM and WB analyzed the data, interpreted the results, and contributed in drafting the manuscript. All authors read and approved the final manuscript.
Ali Mehryar Karim, MBBS, PhD
Senior Monitoring & Evaluation and Research Advisor
The Last Ten Kilometers Project, JSI Research & Training Institute, Inc., Ethiopia
Addis Tamire, MBBS, MPH
Chief of Staff
Ministry of Health, Federal Democratic Republic of Ethiopia
Araya Abrha Medhanyie, MPH, PhD
Faculty, Department of Public Health, College of Health Sciences, Mekelle University, Mekelle, Ethiopia
Wuleta Betemariam, MA, MPH
Director
The Last Ten Kilometers Project, JSI Research & Training Institute, Inc., Ethiopia