Distribution of maternity waiting homes and their correlation with perinatal mortality and direct obstetric complication rates in Ethiopia

Background: Ethiopia has been expanding maternity waiting homes to bridge geographical gaps between health facilities and communities in order to improve access to skilled care. In 2015, the Ministry of Health revised its national guidelines to standardize the rapid expansion of waiting homes. Little has been done to document their distribution, service availability and readiness. This paper addresses these gaps as well as their association with perinatal mortality and obstetric complication rates. Methods: We utilized data from the 2016 national Emergency Obstetric and Newborn Care assessment, a census of 3804 public and private health facilities. Data were collected between May and December 2016 through interviews with health care workers, record reviews, and observation of infrastructure. Descriptive statistics describe the distribution and characteristics of waiting homes and linear regression models examined the correlation between independent variables and institutional perinatal and peripartum outcomes. Results: Nationally, about half of facilities had a waiting home. More than two-thirds of facilities in Amhara and half of the facilities in SNNP and Oromia had a home while the region of Gambella had none. Highly urbanized regions had few homes. Conditions were better among homes at hospitals than at health centers. Finished floors, electricity, water, toilets, and beds with mattresses were available at three (or more) out of four hospital homes. Waiting homes in pastoralist regions were often at a disadvantage. Health facilities with waiting homes had similar or lower rates of perinatal death and direct obstetric complication rates than facilities without a home. The perinatal mortality was 47% lower in hospitals with a home than those without. Similarly, the direct obstetric complication rate was 49% lower at hospitals with a home compared to hospitals without. Conclusions: The findings should inform regional maternal and newborn improvement strategies, indicating gaps in the distribution and conditions, especially in the pastoralist regions. The impact of waiting homes on maternal and perinatal outcomes appear promising and as homes continue to expand, so should efforts to regularly monitor, refine and document their impact.

The country has a decentralized health system with three tiers where the first level provides primary health care and acts as the major platform for health service delivery. It consists of one primary hospital with four or five primary health care units (PHCUs). A PHCU is composed of a health center and five satellite health posts to serve approximately 25 thousand people. Health centers are staffed with health officers, nurses, midwives, and laboratory technicians to provide primarily preventive care including ANC, delivery and post-natal care, curative, inpatient and ambulatory services, including maternal and child health (MCH) services. It serves as a referral center and administrative and technical linkage to health posts. A primary hospital provides inpatient and ambulatory services to an average population of 100,000. It also provides emergency surgical services, including cesarean sections and access to blood transfusion services, and serves as a referral center for the health centers in its catchment area while serving as a practical training center for nurses and other paramedical health professionals. General hospitals provide care at the secondary level to a catchment population of approximately one million people. They serve as referral and training centers for primary hospitals and mid-level professionals. The third level is a specialized hospital that serves a catchment population of about five general hospitals or 5 million people. All public health centers and hospitals, as well as private hospitals and MCH specialty clinics, are expected to provide delivery services. Ethiopia is committed to improving maternal and newborn health outcomes and its targets are aligned with those of the Sustainable Development Goals. To improve outcomes, Ethiopia aims to strengthen health systems to provide universal access to high quality promotive, preventive, curative, and rehabilitative services. This strategy is laid out in the Health Sector Transformation Plan (HSTP 2015–2020). During the HSTP period, the Federal Ministry of Health has developed different strategies and initiatives including the establishment of effective clinical mentorship and quality improvement initiatives. Moreover, the government seeks to improve access to and utilization of EmONC services by promoting facility delivery, expanding MWHs at health centers [17], strengthening referral linkages through the procurement and distribution of ambulances, expanding the number of health facilities, and the number of midwives, emergency surgical officers, and specialty professionals to ensure EmONC services [22]. Maternity waiting homes are expected to be available in most rural health centers which are closer to the rural population than other health facilities. This is a secondary analysis of the 2016 national EmONC assessment [18], a national cross-sectional census of 3804 public and private health facilities that provided maternal and newborn health services. All public hospitals (referral, general, primary) and health centers, and all private (for-profit and not-for-profit) facilities (hospitals, MCH specialty centers, MCH specialty clinics, and higher clinics) that reported having attended births in the 12 months prior to the survey were included in the study. Facilities classified as medium clinics or below were excluded per the guidance of the Food, Medicine and Health Care Administration and Control Authority of Ethiopia, who sets out which facilities are expected to provide childbirth services. The 2008 Ethiopia EmONC assessment modules (questionnaires) and a set of survey tools revised by Columbia University’s Averting Maternal Death and Disability Program (AMDD) in 2014 were adapted to the national context. The Ethiopian Public Health Institute (EPHI) designed an electronic data collection template using CSPro 6.1. Data were collected between May and December 2016 through interviews with health care workers, record reviews, and observation of infrastructure. The overall assessment utilized 14 facility-based modules. Ethiopia was the first country to test the MWH module. It included data related to the infrastructure, support, and features of the MWH as reported by the facility medical director or designee. For this secondary analysis, we used the facility case summary and maternity waiting home modules. Data were managed using CSPro 6.1 programming and exported to Stata 15.1 for statistical analysis [23]. Distribution, infrastructure, and characteristics of MWHs were described and the association between independent and dependent variables were analyzed using univariate and multivariate linear regression models where the unit of analysis was the facility. The outcome variables considered in this analysis were the institutional perinatal death rate (PDR) and direct obstetric complication rate (DOCR) in the 12-month period preceding the assessment. We defined perinatal deaths as all stillbirths (macerated and fresh) and all live births who died within 24 h or before discharge, whichever came first. The perinatal deaths were divided by the number of deliveries that took place in the facility over the same period and multiplied by 100. The DOCR was defined as the proportion of admitted women who had a major obstetric complication (antepartum or postpartum hemorrhage, retained placenta, severe pre-eclampsia or eclampsia, severe abortion complications, uterine rupture, ectopic pregnancy and prolonged/obstructed labor) as well as any other direct obstetric complication (multiple gestation, premature rupture of membranes, etc.). It was calculated as the number of women with obstetric complications treated divided by the number of deliveries recorded in the same facility, multiplied by 100. We performed logarithmic transformations on each outcome variable prior to running the models to achieve a more normal distribution; thus, regression coefficients should be interpreted as percent change. The independent variable of primary interest was the availability of a MWH. Moreover, we included region, managing authority of the facility, location of facility (urban/rural), availability of motor transport, density of skilled birth attendants (SBAs) per annual deliveries, and volume of annual deliveries. Other variables and their operational definitions used in this study are presented below. EmONC facility: EmONC is defined as a set of life-saving interventions used to treat the major obstetric causes of morbidity and mortality. To assess the level of care, the performance of these signal functions in the last 3 months defines whether a facility is classified as providing basic EmONC (BEmONC) or comprehensive EmONC (CEmONC). BEmONC services comprise: 1) administration of parenteral antibiotics to prevent puerperal infection or treat abortion complications; 2) administration of parenteral anticonvulsants for treatment of eclampsia and preeclampsia; 3) administration of parenteral uterotonic drugs for postpartum hemorrhage; 4) manual removal of the placenta; 5) assisted vaginal delivery (vacuum extractions); 6) removal of retained products of conception; and 7) neonatal resuscitation with bag and mask. CEmONC services comprise cesarean sections and blood transfusions, in addition to all BEmONC functions [13]. Index of MWH infrastructure and amenities: an index score was calculated for each MWH, measured by 10 infrastructure and amenity indicators listed in Table 1. Each item was given a score of 0–2 points: 2 for each item available that met the standard, 1 for partial availability and 0 for not available or below the standard. Three items had a maximum of 1 point. Items were given equal weights and a total score was generated; maximum score being 17. Waiting homes that scored 13 or more were categorized as optimal, scores in the range of 9–12 points were ranked as basic, and scores less than 9 were considered substandard. Items used to measure MHW index score

N/A

• Dima Health