- Top of page
- CONFLICT OF INTEREST
Ethiopia has great needs in the area of maternal and newborn health. The most recent maternal mortality ratio estimates for 2010 and 2011 range from 350 deaths to 676 deaths per 100,000 live births, depending on the source. Nearly half of pregnant women received antenatal care, yet only 10% of women gave birth with a skilled provider in attendance; more than 61% of women indicated that a health facility is not necessary for birth; and 7% received postnatal care within 2 days of birth. While Ethiopia has made considerable progress in improving child survival between 2005 and 2011, the neonatal mortality rate only decreased from 39 to 37 deaths per 1000 live births.[2, 3] Most maternal and newborn deaths occur at birth or within the early postnatal period.[4, 5] Rural residence, timing of maternal and neonatal mortality, and limited access to skilled providers and health services create challenges and opportunities for improving maternal and newborn health outcomes using community-based strategies.[6-8]
In this article, we report the main findings of an evaluation of the Maternal Health in Ethiopia Partnership (MaNHEP), a project to develop a community-based model of maternal and newborn health in rural Ethiopia and to position it for scale-up. We describe the extent to which the project's objectives were met and highlight implications for policy, programming, and research. Other articles in this supplement issue of the Journal of Midwifery & Women's Health provide more detailed descriptions and results of MaNHEP interventions.[10-17]
Maternal and Newborn Health in Ethiopia Partnership
MaNHEP was a 3.5-year (November 2009-May 2013) learning project funded by the Bill and Melinda Gates Foundation. The project operated under the leadership of the Ethiopian Federal Ministry of Health and 2 regional health bureaus and was aligned with 3 key national policies: Health Sector Development Plan IV Ethiopian Fiscal Year 2004 (equivalent to 2011-2012); National Reproductive Health Strategy 2006-2015; and Road Map for Accelerating the Attainment of the Millennium Development Goals Related to Maternal and Newborn Health in Ethiopia 2011-2015. MaNHEP complemented and strengthened the Federal Ministry of Health's flagship Health Extension Program, which seeks to expand health care delivery to rural areas where 85% of Ethiopians reside by ensuring the delivery of a core package of community maternal and newborn health (CMNH) care (Table 1) to achieve Millennium Development Goals 4 and 5 to reduce child and maternal mortality. MaNHEP sought to strengthen district-wide administrative and health systems to support and continuously improve CMNH care. MaNHEP was led by Emory University, Atlanta, Georgia, in collaboration with JSI Research & Training Institute, Inc., Boston, MA; University Research Company LLC Bethesda, MD, and Addis Ababa University, Ethiopia. MaNHEP was implemented in the Amhara and Oromiya regions at the request of the Federal Ministry of Health.
The Maternal and Newborn Health in Ethiopia Partnership's (MaNHEP's) community-based model focused on maternal and newborn survival during the vulnerable birth and early postnatal period.
The model improved capacity and confidence of health extension workers, community health development agents, and traditional birth attendants to provide maternal and newborn health care during birth and the early postnatal period.
The model increased demand for skilled maternal and newborn care and improved the self-care behaviors of women during labor and the early postnatal period.
The model's lead woreda (district) approach improved identification of pregnant women, enrollment of pregnant women in antenatal care and in MaNHEP's Community Maternal and Newborn Health family meetings, labor and birth notification to health extension workers, and the timely postnatal care follow-up by health extension workers.
Application of the model is associated with improved perinatal survival, as evidenced by an increase in the number of days between perinatal deaths over the course of the project.
Table 1. Community Maternal and Newborn Health Care Package
|Care Before Birth|
|Birth preparedness and complication readiness|
|Promotion of antenatal care|
|Promotion of CMNH family meeting|
|Care at Birth|
|Uterotonics (misoprostol [Cytotec])|
|Care After Birth|
|Trauma check (fistula)|
|Nutrition (especially fluids)|
|Illness recognition and care-seeking|
|Care at Birth|
|Immediate care (dry, stimulate, keep warm, delay cord clamping)|
|Care After Birth|
|Promotion of immediate, exclusive|
|Thermal care, kangaroo mother care|
|Clean cord care|
|Illness recognition and care seeking|
|Case management for pneumonia (where part of national policy)|
The Ethiopian Health Extension Program is implemented through health extension workers, who are paid government employees, primarily young women. They have at least a 10th grade education and one year of training on a package of primary health care focusing on preventive and selected curative heath, including maternal and newborn health. Since 2005, the government has trained and deployed more than 33,000 health extension workers to cover rural communities at a ratio of one health extension worker per 2500 persons.[21, 22] Community outreach was supplemented through a network of community health development agents, who operated at a ratio of one per 30 to 50 households. These agents are women and men, who are somewhat older and have more years of education than the health extension workers. MaNHEP focused its efforts on health extension workers and community health development agents and on traditional birth attendants (TBAs) who, in Ethiopia, are older and have less formal education than the other 2 provider groups. The TBAs assist women at birth and during the early postnatal period. While some TBAs learned basic midwifery through short-term biomedical training, most acquire their skills through apprenticeship and experience.
The MaNHEP project design rested on 4 assumptions. First, health extension workers are not likely to be present at birth. Family members and TBAs are most likely to provide care. Second, opportunity exists for entry of health extension workers into the home during birth and during the early postnatal period, in addition to or in place of TBAs or other less skilled caregivers (eg, untrained family members). Third, this shift will be made more easily by establishing a team comprised of health extension workers, community health development agents, and TBAs—and by introducing a CMNH care package (Table 1) through these care providers. Fourth, an effective team can provide a platform for delivering additional health interventions at the local level.
The project's aim was to learn how best to ensure that the CMNH care package could be provided to “every woman, in time, every time.” Toward this aim, MaNHEP pursued 3 objectives: 1) improve the capacity and performance of the team of health extension workers, community health development agents, and TBAs to provide the targeted CMNH care; 2) increase demand for the targeted CMNH care and improve self-care behaviors; and 3) develop and demonstrate the effectiveness of lead woredas (districts) to improve CMNH care and services. Lead woredas are model districts that are committed and able to continuously improve care and service delivery to meet the needs of childbearing families.
MaNHEP deployed a 3-pronged intervention to achieve its objectives and to influence both supply and demand for CMNH: 1) a CMNH training program, 2) continuous quality improvement, and 3) behavior change communications. The first intervention component, a community training program, aimed at objectives 1 and 2, worked with district health and community stakeholders to teach the CMNH care package. The training program, adapted from the American College of Nurse-Midwives’ (ACNM's) Home Based Life Saving Skills program, had 2 components. In a facility-based component, health extension workers were given refresher clinical training for safe, clean birth and postnatal care to build competence and confidence in providing the targeted care. In a community-based component, health extension workers, community health development agents, and TBAs shared their knowledge and expertise in week-long training sessions, acquiring new knowledge and skills to provide better care. After training, community health development agents and TBAs worked in pairs called “guide teams” to teach skills (eg, clean birth, safe use of misoprostol [Cytotec], uterine massage to reduce postpartum bleeding after delivery of the placenta, and newborn resuscitation) to women in their second and third trimester of pregnancy and their family caregivers, who will be present at birth, during community CMNH family meetings.
The second intervention component, continuous quality improvement, was aimed at objectives 2 and 3.[25, 26] Quality improvement teams were comprised of community stakeholders including kebele (the smallest administrative unit of Ethiopia, similar to a village or neighborhood) council members, priests, agricultural association and women's association leaders, health extension workers, community health development agents, and TBAs. The quality improvement teams used participatory problem solving to generate and test ideas to improve pregnancy identification, antenatal care registration, CMNH family meeting attendance, labor and birth notification, and postnatal care follow-up within 48 hours of birth by a health extension worker. The quality improvement teams engaged in iterative cycles of idea generation and testing and met together 5 times over the course of the project to share their learning. The most successful ideas or solutions across the project's 51 kebeles, considered the best care-delivery processes, were compiled in a CMNH Change Package. Intervention components 1 and 2 were designed to collectively build the capacity of the woredas to become lead woredas.
The third intervention component, behavior change communications, reinforced objectives 1 and 2. Through culturally appropriate, professionally produced films, radio and TV dramas, songs, and poetry contests, the project sought to influence community demand for improved care and services and to promote teamwork among the health extension workers, community health development agents, and TBAs for better service delivery.
- Top of page
- CONFLICT OF INTEREST
MaNHEP was implemented in 3 woredas (districts) in each of the 2 regions: North Achefer, South Achefer, and Mecha in the Amhara region, and Degem, Kuyu, and Warejarso in the Oromiya region. Results are presented by region. Woreda selection was based on need; population size and number of expected births; accessibility of health services; presence of health extension workers, community health development agents, and TBAs; and the absence of other development partners working on the same or similar issues. MaNHEP focused on a section of each woreda consisting of 2 health centers and their associated health posts and all pregnant or newly postpartum women and newborns within the health posts’ catchment areas. Overall, MaNHEP implementation encompassed a population of about 350,000 residing in 51 kebeles, with about 12,000 births per year.
MaNHEP used an uncontrolled before/after study design, with data for evaluating project level indicators primarily collected from randomly selected respondents through surveys and skills assessments, as well as a quality improvement survey and a verbal autopsy study with a one-year open cohort of pregnant women. Monthly quality improvement monitoring complemented these methods and permitted a degree of data triangulation.
The results for objective 1 indicators are based on data obtained from a health care provider baseline and endline survey[28, 29] and data obtained from clinical skills assessment. The baseline survey was conducted from June through September 2010, and the endline survey was conducted from May through August 2012. The survey questionnaire, which focused on CMNH knowledge, practice, and coverage, was developed in English, translated to Amharic and Oromifa, back-translated, pretested, and refined before use. It was conducted in the local language by MaNHEP-trained interviewers and involved individual face-to-face interviews with a stratified random sample of 266 (baseline) and 264 (endline) health extension workers, community health development agents, and TBAs. Individuals surveyed were asked which of the CMNH care elements they had provided at the last birth attended (yes or no). The proportion of care elements reported as having been provided, out of a total of 17 care elements, was calculated to measure individual completeness of care. The values were averaged across the individuals surveyed. Individuals surveyed were also asked to self-assess their level of confidence in the ability to provide pregnancy, birth, postpartum, and newborn care on a Likert-type scale (1 = not very confident, 10 = very confident). The values were averaged across the individuals surveyed. Finally, individuals surveyed were asked to self-assess whether they viewed themselves as part of a CMNH care team (yes or no). The proportion of individuals who saw themselves as part of a team, out of the individuals surveyed, was calculated.
The clinical skills assessment focused on labor, birth, and early postnatal care. The assessment was conducted by trained MaNHEP CMNH specialists and involved direct observation using skills checklists. Each health extension worker, community health development agent, and TBA was given a scenario and the appropriate props and was asked to demonstrate how to care for a woman and newborn (eg, helping a newborn breathe using a special resuscitation doll).The skills checklists, part of ACNM's Home-Based Life Saving Skills program, have been used successfully in more than 15 countries, including in the initial field tests of that program in Ethiopia from 2001 to 2004. The baseline pre- and immediate posttraining skills assessment included all 808 health extension workers, community health development agents, and TBAs trained by MaNHEP. The endline postintervention assessment included a stratified random sample of 359 of these workers. Individuals were tested to determine if they could correctly demonstrate the CMNH care steps. The proportion of CMNH care steps correctly demonstrated, among all the care steps in the skills checklist, was calculated for each individual observed to yield a test score. These scores were averaged across the individuals tested.
The results for objective 2 indicators are based on information obtained from a women's baseline and endline survey.[28, 29] This survey was comparable to the health care provider survey in content and was conducted by the MaNHEP-trained interviewers during the same time frame. The baseline and endline surveys involved face-to-face interviews with systematic random samples of 1027 and 1019 women, respectively, who gave birth during the year before the survey. Each woman surveyed was asked whether she knew the health extension workers, community health development agents, and TBAs in her own kebele (yes or no). The proportion of individuals who knew each provider group, out of the individuals surveyed, was calculated. Each woman surveyed was asked whether she had received each of 17 CMNH care elements at her last birth (yes or no). The proportion of care elements reported as having been received, out of the 17 care elements, was calculated to measure individual completeness of care. These values were averaged across the women surveyed. Each woman surveyed was asked to self-assess her level of trust in each group to provide CMNH care on a Likert-type scale (1 = least trust, 5 = most trust). The values were averaged across the individuals surveyed. Finally, each woman surveyed was asked whether she had received antenatal care, birth care, and postnatal care (yes or no). For each care component received, she was asked who had provided the care (skilled provider [physician, midwife, or nurse], health extension worker, community health development agent, TBA, unskilled family, friend, or no one), when she received the care (time in relation to pregnancy or birth), and in the case of antenatal care, how often (the number of visits).
The results for objective 3 indicators are derived from data obtained through a variety of sources. The first source was an endline quality improvement survey conducted in October 2012. The CMNH Change Package of successful solutions provided the content of the survey. This survey questionnaire was translated, back-translated, pretested, and finalized before use. A MaNHEP-trained interviewer administered the survey to 51 original quality-improvement teams and 39 new quality-improvement teams, one team per kebele. For each area of emphasis, quality improvement teams surveyed were asked if their kebele had adopted each improvement idea or solution (yes or no). The proportion of solutions that each kebele adopted, out of the number of successful solutions, was calculated. These values were averaged across 51 original kebeles and 39 new kebeles.
The second data source was a verbal autopsy study conducted by a team of trained interviewers from Addis Ababa University. Verbal autopsy is a research method that uses interviews to determine probable social and medical cause of death when an autopsy is not done, medical records are unavailable or nonexistent, or the person who died received no medical attention. In this study, the mothers and others present at birth (if available) were interviewed using the World Health Organization's 2007 standard verbal autopsy questionnaire, an instrument that had been adapted by others for use in Ethiopia. Relevant to objective 3, this team identified and enrolled a cohort of approximately 9500 pregnant women from March 1, 2011, through February 28, 2012, and followed them through birth. The verbal autopsy data included the dates of birth and death for 175 perinatal deaths among this cohort.
For survey and clinical skills assessment data, simple univariate statistics were used to describe each indicator (proportion, mean, and standard deviation), and bivariate statistical tests, appropriate to level of measurement, to assess regional differences in the indicators. To measure the impact of the program on perinatal mortality (stillbirths and early neonatal deaths), the change in mortality pattern was assessed and documented using a quality assurance analysis referred to as a statistical control process analysis, specifically a G-chart that is designed to produce an estimate of units that occur between rare events or of nonconforming incidents (in this case, the number of days between perinatal deaths).[33, 34] The G-chart presents the data in terms of mean, standard deviation (sigma 1 and 2), and upper control limit and gives a picture of deaths over time.
- Top of page
- CONFLICT OF INTEREST
MaNHEP's theory of action proposed that increasing the provision of essential CMNH care in and around the time of birth through improved interactions (more, better, and more equitable) between the team of health extension workers, community health development agents, and TBAs and childbearing women and families would contribute to increased maternal and newborn health and well-being, as well as to reduced mortality and morbidity. Compared with the project's formative research and baseline survey, which together revealed a very limited role and involvement of health extension workers, community health development agents, and TBAs in CMNH care, the observed improvements by endline are striking.
The first project objective—improve the capacity and performance of a team of health extension workers, community health development agents, and TBAs to provide targeted maternal and newborn health services—was met. The project recorded substantial improvements in the completeness of the CMNH care that they provided at the most recent birth attended, in both their demonstrated ability and self-reported confidence to provide this care, and in their sense of being part of a care provider team.
The second objective—increase the demand for targeted CMNH services and improved self-care practices—was also met. There were significant improvements in women's awareness of and trust in the health extension workers’, community health development agents’, and TBAs’ ability to provide care and in the completeness of care that women reported receiving from these providers at their most recent birth. Moreover, women's use of skilled providers and health extension workers for antenatal and postnatal care services and timing of these services improved substantially. There was a shift toward the use of providers with a higher level of skills for birth care, but this shift differed in the 2 regions. There were increases in Amhara women's use of skilled providers (7%-18%), health extension workers (3%-19%), and health facilities (6%-17%), but there were no comparable increases in Oromiya. In both regions, however, there were increases in the use of trained TBAs, coupled with substantial decreases in reliance on family members and other unskilled providers for birth care. These improvements in women's use of skilled providers and health extension workers are very encouraging and in line with national priorities and policies, particularly the engagement of health extension workers in ambulatory care.[19-21] With respect to maternal health, there is a clear need for continued improvement, particularly with regard to the use of skilled providers and an appropriate use of health facilities for normal birth and emergency obstetric care. For rural communities in Ethiopia, this transition to facility-based care will take time.
The third objective—demonstrate the effectiveness of lead woredas—was met. The large majority of successful local solutions to improve pregnancy identification, antenatal care registration, labor and birth notification, and postnatal care follow-up within 48 hours were adopted by the original and new project kebeles. The significant increase in number of days between perinatal deaths in the context of increased exposure to the project interventions also suggests a positive association between key intervention components (CMNH family meetings and quality improvement activities) and improved perinatal outcomes. The role of the behavior change component in this result is unclear, given a later implementation. Finally, discussions of potential regional scale-up are being held.
MaNHEP was a learning project, implemented in woredas where there were no other organizations with projects focusing on CMNH. Evaluation of the project objectives was based on an uncontrolled before/after study design, with data for project-level indicators collected through baseline and endline knowledge, practice, coverage surveys and clinical skills assessments,[28, 29] a quality improvement survey, and a verbal autopsy study. While the uncontrolled before/after design is appropriate for a learning project, the limitations of this design are well known and include a variety of threats to internal validity. This design can be used to determine the magnitude and direction of a change, but it does not permit attribution to some or all of the components or to external factors. As with all self-report methods, the survey and verbal autopsy data are subject to recall and social desirability biases. However, the baseline and endline surveys[28, 29] anchored the monthly quality-improvement monitoring that tracked changes in selected indicators, including CMNH family meeting attendance, antenatal care registration, labor-birth notification, and postnatal care follow-up over the course of the project before the endline survey. This monthly monitoring also documented the solutions that were tested by the quality improvement teams and that were associated with the observed changes. MaNHEP also facilitated and documented regularly occurring learning sessions and coaching visits. Finally, MaNHEP tracked completeness of CMNH care through quarterly interviews with random samples of 10% of women who gave birth in the previous 3 months and whose births were reported to the health extension workers. The change observed between baseline and endline surveys was consistent with the change observed in monthly quality-improvement monitoring and quarterly birth audits. This consistency increases confidence in the findings.
Several features of MaNHEP's model were essential to its success. First were the partnership, leadership, and active engagement of the Ministry of Health and local communities. Regional health bureaus and zonal health departments provided political leadership and support, while a variety of Woreda health office, health center, and health post leaders and staff actively participated in the CMNH training as well as quality improvement training, learning sessions, and monthly coaching of guide teams and quality improvement teams. Moreover, the project was implemented within a national policy environment that prioritized achievement of Millennium Development Goals 4 and 5, creating a shared vision that enhanced the partnership.
Second, the project operated from focused, mutually reinforcing objectives and skills-based intervention (CMNH clinical training and CMNH family meetings, continuous quality improvement, and behavior change communications) that targeted and linked pregnant women and those who will be present at birth, relevant community groups, and the health sector at different levels. The model was tightly integrated horizontally (through activities) and vertically (through actors) to improve care during and around the time of birth, and it included built-in mechanisms for ongoing support and supervision. MaNHEP's CMNH content is similar to other community-based programs that aim to reduce maternal and/or newborn morbidity and mortality and improve health outcomes. However, the model of integration differs from intervention packages that have been implemented by a range of community health workers, such as programs that consist primarily of building community support groups/women's groups; community mobilization, antenatal, and postnatal home visits; antenatal care and TBA training for home birth care; community mobilization and home-based newborn care; or home-based newborn care by itself.
Third, the project emphasized continuous, collaborative learning. While the desired general outcome of each component intervention was predetermined, the processes for achieving each outcome were not. Those processes evolved through an organic development of local ideas about how best to ensure that women and their newborns receive the targeted care. The project created a space for dialogue, development of trust, change, and local ownership.
Finally, MaNHEP demonstrated an ability to adapt to a changing policy landscape. Midway through implementation, the Federal Ministry of Health dramatically modified the Health Extension Program structure. In 2012, a volunteer health development army was launched with the goal of creating a network structure ensuring one volunteer per 5 households. The health development army is supervised by health development army team leaders (also part of the network) who, in turn, are supervised by woreda health center specialists (for administrative support) and health extension workers (for technical support). This new structure is in various stages of implementation nationally. Because MaNHEP's structure is similar to the new health development army structure (in its sub-kebele level guide teams, kebele-level quality improvement teams, and work with the chain of government health agencies), adaptation and integration is occurring in new kebeles within the project area woredas in a relatively seamless manner at the request of the regional health bureaus. MaNHEP's model is adaptable and potentially scalable.
Future program and research efforts might include the addition and evaluation of other primary health care components to the model. A detailed qualitative study of factors associated with the differential development of motivation, self-efficacy, and empowerment of those participating in the project interventions, including woreda administration, quality improvement coaches, health extension workers, community health development agents, TBAs, women, and families, would be helpful. Insights on these more intangible benefits of participation and their sustainability would help to inform interventions that strengthen the health extension program and primary health care unit structure in Ethiopia. Given MaNHEP's achievements, we suggest that next steps include a high-quality cluster randomized controlled trial that includes cost-effectiveness. This is important for Health Extension Program policy and programming in Ethiopia and would contribute to the growing body of research on the effectiveness of CMNH care.