Most hospitals had continuous electricity and water supply, which contrasts with previous studies (13, 43–45). However, gaps were highlighted in the availability of telecommunication infrastructure, with less than half of the hospitals having had telephone service for internal communication, and only one hospital had the same in the compound for use by patients and their families. A study in Northern Ethiopia has reported that fewer facilities (12.5%) had a working phone or shortwave radio (43). This compares to 100% of hospitals in Zambia and 88% in Uganda having functional communication systems (44). However, this finding does not take into consideration the use of cell phones. This profound and technological change in telecommunications has not been updated within the Ministry of Health formal audit. However, cellphones are personal rather than institutional raising questions on payment and access. Functional communication systems facilitate communication among health facilities, contributing to improved referral networks for women with complications (44). It also contributes to a positive experience of women with the care process, through improving responsiveness to their needs such as access to information and social and emotional support (46).
Availability of transport is important to ensure that women with complications are able to reach facilities that can provide appropriate care. A functional ambulance for emergency referral transportation was available in a great majority (90%) of the hospitals. This is much higher than the figures reported in previous studies in the region, including 65% (13), 59.4% (43), and 61% (44). This may be partly explained by a difference in the definition of the indicator. This study measured whether a functioning ambulance or other vehicle was stationed at the hospitals for emergency transport other studies have included an additional criterion of whether the vehicle had fuel available on the day of the survey (13), while others measured transport across a range of services including health centres and government, private and non-government managed hospitals. As financing and support systems are important features of operational environment affect facility performance, differences in supply chains and authorization are important. Government facilities depend on district and provincial administration for salary and material inputs (38).
Previous studies showed that women have limited opportunities to express opinions about their childbirth experience, or log complaints due to lack suggestion boxes or a log book (19). Our study found that most hospitals had a suggestion box or log book in the labour ward for handling complaints. However, only half the suggestions were regularly evaluated, and documented. The lack of response may indicate a lack of accountability, and an undermining of women’s rights to seek justice for their mistreatment or malpractice (19).
The Ethiopian Ministry of Health requires that all health facilities have a written, up-to-date policy on triage and waiting times for emergency and non-emergency consultations and treatment for maternal and newborn healthcare (18). However, one third of the hospitals did not have an emergency triage system for sick pregnant mothers who were not in labour. This is another gap that needs improvement.
Obstetric services must always be available, as an obstetric emergency can happen at any time (39). One or more skilled attendants were present in all hospitals including weekends, for routine L&D care. Almost all hospitals had four or more fulltime midwives, though half the hospitals lacked a fulltime specialist in gynecology and obstetrics or a general practitioner. This compares favorably to 53.1% in Northern Ethiopia (44). The recent SPA + Survey (13) and a facility survey in Uganda and Zambia (44) provided similar observations.
Yigzaw et al (48) have identified lack of training among midwives as a barrier for provision of quality L&D care. Limited opportunities for refresher training on obstetric care may indicate less emphasis given to the importance of in-service training in improving the quality of L&D care (49). Ongoing practice and periodic refresher training maintains obstetric management skills (47). Skilled attendants were trained on management of L&D in the 12 months prior to this study in half of the hospitals, (similar to Northern Ethiopia (56.3%) (43).
The required essential drugs, supplies and equipment should be kept in the L&D ward at all times to avoid unacceptable delay in providing the services (39). This study showed that all essential drugs and equipment were available in only a quarter of the hospitals, and even fewer (15%) hospitals had all the supplies. However, oxytocin, an effective intervention to prevent postpartum hemorrhage was universally available, while magnesium sulphate, an effective intervention to manage pre/eclampsia, was missing in two hospitals.
Previous studies have also highlighted gaps in meeting the standards for essential drugs, supplies and equipment, though the degree of availability varies. For example, Getachew et al(49) reported that 52.6% of the hospitals in Ethiopia (N = 19) had all the medicines and supplies needed for normal delivery, and 16% had magnesium sulfate. Centers for Disease Control and Prevention(44) also reported universal availability of oxytocin and magnesium sulfate in Zambian hospitals. Both studies documented higher availability of drugs than the present study (except for oxytocin) which may be partly explained by methodological and contextual differences between the studies. Hospitals from different managing authorities (government and missionary) across the country were included in the former study, while contextual difference and inclusion of different levels of facilities were in the latter study.
Shortage of HIV test kits and ARVs is a structural barrier faced by many hospitals. Official policy in Ethiopia is that all women with unknown HIV status should receive a rapid HIV test in maternity wards so that if diagnosed positive, ARV drugs can be given to the mother and baby in time to prevent vertical transmission (39). Two hospitals lacked HIV test kits and ARV drugs for HIV positive mothers and for exposed babies. This situation in Ethiopia is similar to other African countries. In Zambia, HIV test kits were universally available and 50% of hospitals had ARVs drugs. In Uganda, 69% of the hospitals had ARV drugs (44). To stop new vertical transmission of HIV, the coverage of Prevention of Maternal to Child Transmission (PMTCT) of HIV service should be above 85%. Yet, only 60.6% of HIV-positive Ethiopian women receive ART (50).
Notably, four in ten hospitals lacked towels for drying and wrapping newborns which highlights a critical gap in the hospitals’ capacity to promote thermoregulation in newborns. This is in agreement with previous work showing that this essential relatively inexpensive intervention had not been properly instigated. De Graft-Johnson et al (47) identified this as the largest gap in the supply and Fisseha et al (43) reported that the supply was available in 15.6% of the facilities.
Quality L&D care involves preventing avoidable infections thus the availability and accessibility of infection prevention IP equipment and supplies is essential in all facilities to enable health workers to adhere to the recommended hygiene practices (18). Almost all hospitals had consumables for IP and adequate safety boxes and color-coded bins for waste segregation. Personal protective equipment was also fulfilled in most hospitals. This is better than a similar study which reported that the required personal protective items were available in two third of the facilities (43).
Offering quality labour and delivery care also involves a welcoming and clean environment including access to a reliable supply of safe water and toilet facilities in maternity ward (39). Almost all hospitals had well ventilated rooms, and the majority had tap water, sufficient space, a functional and clean toilet, and functional sinks with detergents. Availability of safe water supply is comparable with a study that reported 65.6% (43), but lower than the Centers for Disease Control and Prevention (CDC)(44) surveys which reported all hospitals in Uganda and Zambia having suitable supplies. This difference could be due to definitional criteria, with CDC using, regular water supply versus the specific definition used in this study, i.e., regular tap water supply.
The availability of toilet facility is favored by the Ethiopia SPA (+)2014 Survey which reported an average of 74% of facilities having had a functioning latrine facility (13), but much higher than 34.4% reported in Northern Ethiopia (43). Again, difference in the definition of the variable is the likely explanation for the deviation. The Northern Ethiopia study (43) measured the variable as availability of functional toilet and shower, while the current data did not specify the ‘shower’ element.
The capacity to conduct laboratory tests significantly enhances the quality of L&D care services. These require continuous supply of safe blood for emergency transfusions to treat hemorrhage (13). However, critical gaps were identified with only 15% of hospitals having the range of tests needed for a mother in labour. This is consistent with other studies including the Ethiopia SPA (+)2014 Survey (13), and those in Malawi with low scores for laboratory system (45).
The results of this study showed that 85% of hospitals received blood from blood banks with safe storage practice, which is better than 75% in Ugandan hospitals (44). This encouraging finding may reflect leadership commitment and efforts made to improve availability of safe blood and blood products in hospitals across Ethiopia. For instance, the existing blood banks have been reorganized by transferring the management of blood transfusion services from the Ethiopian Red Cross Society to government, the National Blood Transfusion Service. New blood banks are also constructed (50). Due to lifesaving nature of the service, efforts should however be strengthened further to close the observed gap, even if small (15%).
All hospitals should display protocols and guidelines on MNHC for staff (39). However, only half of the hospitals had all essential guidelines and protocols in the L&D ward. The frequently missing guidelines for normal birth were management protocol on selected obstetrics topics (2010 version) and infection prevention protocol. Previous studies have also reported this poor performance (45, 49). Accessibility to the best evidence determines successful implementation of evidence-based medicine (20), for example, the use of printed job aids that provide prompts to remind providers to perform specific tasks during intrapartum care is the most important predictor of quality of EmOC (51). Thus, the result indicates another opportunity for QI.
Maternal and neonatal death review is important for improving the quality of MNHC services through systematic process of identifying factors associated with the deaths, generating recommendations to develop interventions against future similar deaths, and measuring improvement (39). Almost all of the 20 hospitals conducted monthly maternal and neonatal death audits and implemented the recommendations. On the other hand, despite two thirds of the hospitals having a MNHC QI committee with a coordinator or focal person assigned, seven of the committees were not conducting regular meetings. This may be due to a lack of an organizational framework and standard procedures indicated in the EHAQ package (39). Some hospitals were conducting death audits in the absence of a MNHC QI committee.
Inadequate performance of health facilities on various QA activities has also been observed in previous studies in Malawi (45) and Ethiopia (13). For example, Ethiopian Public Health Institute (EPHI), FMoH of Ethiopia and ICF International (13) showed that half of all hospitals in Ethiopia had regular QA activities with observed documentation. Effective QA audit and feedback system in place can determine successful implementation of evidence-based practice (20). There is a need to improve the ability to learn lesson from case reviews.
In summary, capacity gaps were observed in the hospitals to provide quality routine L&D care services with about two thirds of the required resources fulfilled overall. Only two hospitals had fulfilled almost all the standards, while one third of the hospitals had low readiness to deliver the service. Laboratory services and safe blood, and essential drugs, supplies and equipment were the areas with the largest gaps. These current findings are generally consistent with a recent study conducted in Northern Ethiopia on quality of delivery service where 65.62% of the facilities had good input quality (43), but much better than another study on quality of midwifery care in Amhara Region, Ethiopia, which reported 16.3% for availability of all essential drugs, and less than 10% for all essential equipment, all supplies, and all IP materials, each (48). The present finding is also favored by other studies. For example, essential equipment for obstetric care was not always available in Kenyan hospitals (17) and health facilities in six sub-Saharan African countries including Ethiopia (47). It is important to note that the results show just a single time point prevalence of the availability of the required inputs for the provision of quality care. This was the best possible assessment given the resources. Thus, they may not necessarily indicate a constant supply.
Gaps in structural quality of L&D care can demotivate providers and constrain their adherence to the actual care practices (13, 15), although quality services can still be provided in minimal service delivery settings (13). Poor facility readiness can also have negative effect on women’s perceptions and service use as reported in Nigeria (16), ultimately weakening the link between health facilities and community (17).
The following limitations are acknowledged in this study. First, the results have limited generalizability to all public hospitals in the country, as the study was conducted only in SNNPR due to resource limitations. The quality of L&D care provided in the public hospitals of the region might differ from other regions. The use of locally adapted EHAQ standards could also compromise the external validity of the study. Representation of facilities from different managing authorities (government, private and non-government) is important as it implies differences in resource allocation, service requirements and working conditions, which can affect the quality of care they provide in different ways (2). Thus, the results from this study may not even reflect the quality of L&D services provision in SNNPR, as it did not include public health centres and private and NGO owned facilities that provide skilled delivery care.
Second, composite scores are averages which can result in loss of important information (52). The results were summarized using composite scores with apparent loss of item specific details. While acknowledging the limitation, composite measures have several advantages including: increased reliability than individual items(52) and giving the picture of both the whole and individual parts (10, 53). This facilitates comparison and simplifies communication of performance on technical components of care to non-technical audiences (53).
Third, an extension of the above limitation relates to the use of unweighted scoring or equal weighting approach for the composite measures. It is acknowledged that assuming a linear relationship between all items and that they are equally important for the provision of quality L&D care might have impacted the results. The use of a weighted scoring was ruled out due to lack of evidence on the relative importance of the constructs of each quality component, though well-defined scientific evidence exists on the constructs, i.e., structural aspects of the EHAQ change package for L&D care (39). Fourth, the availability of equipment, and essential drugs and supplies was based on point prevalence at the time of observation and does not necessarily indicate a constant supply, but this was the best possible assessment given the resources.
Fifth, the data captured on the availability of telephone service might not be valid as the facility audit checklist considered only fixed telephone service. However, due to the profound technological change in telecommunications in Ethiopia, the coverage of cell phones is wider than the landline telephone network. Finally, information on refresher training was captured at hospital level through facility audit. The information might be less valid as it depended on the L&D case team leader’s verbal report which is subject to recall bias. Interviews with selected health professionals would rather provide a more valid measure of the indicator.