Health system readiness to support facilities for care of preterm, low birth weight, and sick newborns in Ethiopia: a qualitative assessment

Background Preterm birth is a worldwide challenge with the highest burden in low- and middle-income countries. Despite availability of low-cost interventions to decrease mortality of preterm, low birth weight, and sick newborns, these interventions are not well integrated in the health systems of low- and middle-income countries. The aim of this study was to assess, from the perspective of key stakeholders comprising leaders in the public health system, the health system readiness to support health care facilities in the care provided to preterm, low birth weight, and sick newborns in different regions of Ethiopia. Methods A qualitative assessment using in-depth interviews with health facility leaders was conducted in health facilities in 3 regions of Ethiopia from December 2017 to February 2018. The interview guide was developed using a modified version of the World Health Organization health system building blocks. Results Across the public health system, adequate and reliable space, power, and water were problematic. Human resource issues (training, staffing, and retention) were critical to being able to properly care for preterm, low birth weight, and sick newborns. Problems with functional equipment and equipment distribution systems were widespread. Funds were lacking to support preterm, low birth weight, and sick newborn needs in facilities. Data collection practices, data quality, and data utilization were all problematic. There were gaps in the availability of guidelines and protocols, specifically targeting preterm, low birth weight, and sick newborn care. Key facilitators, information disseminators, and influencers identified in the study were the Health Development Army, community and religious leaders, and mothers and families who had had positive experiences or outcomes of care. Conclusions The Ethiopian health system has opportunities across all 7 World Health Organization health system building blocks to strengthen readiness to support health facilities to provide quality care and improve outcomes for preterm, low birth weight, and sick newborns.


Background
It is estimated that 15 million (11.1%) babies born worldwide in 2010 were preterm (born alive before 37 weeks gestation) [1]. Preterm birth is a worldwide challenge, with the highest burden in low-and middle-income countries (LMICs). Despite the availability of low-cost interventions to reduce mortality of preterm, low birth weight (LBW), and sick newborns, these interventions are not well integrated in the health systems of LMICs. Neonatal deaths are the leading cause of under-5 mortality [2], and complications of prematurity is the leading cause of neonatal mortality [3,4].
Through substantial investments in public-sector health systems, Ethiopia has achieved notable progress over the last decade in preventing postnatal newborn deaths. Newborns benefit from high-impact interventions that include integrated community case management, community-based newborn care, and prevention of mother-to-child transmission of HIV through Option B+. However, reductions in neonatal mortality have lagged behind reductions in child mortality [4]. The country has one of the highest neonatal mortality rates globally, with 28 neonatal deaths per 1000 live births, and has the sixth highest number of neonatal deaths, with an estimated 90,000 deaths in 2016 [5]. Of under-5 deaths in Ethiopia, 44% occur within the first 28 days of life, a rate that has remained stagnant. The most common causes of death among newborns in Ethiopia are prematurity (37%), infection (28%), and asphyxia or other intrapartum events (24%) [6].
Countries with high neonatal mortality have bottlenecks in almost all of the health system building blocks defined by the World Health Organization (WHO): service delivery, health workforce, health management information systems, access to essential medicines, financing, leadership and governance, and community engagement [7,8]. A study from 12 high-burden countries showed that the most prevalent bottlenecks were in the areas of health workforce and health financing (10 out of 12 countries), followed by community ownership and partnership (9 out of 12 countries) [9].
Establishment of neonatal intensive care units (NICUs) and diversified levels of care have increased the survival of preterm, LBW, and sick newborns less than 32 weeks of gestation, although this may not apply to most countries with the highest burden of preterm deliveries and deaths [10]. Despite the absence of these high-level centers, moderately preterm newborns can be saved with feasible, cost-effective care [11]. Barriers to neonatal care include shortages of qualified personnel, geographical inaccessibility, inadequate regulation of pharmaceutical and private sector providers, insufficient provider knowledge and skills, poor demand for interventions, and lack of equipment and infrastructure. Continuous and multifaceted improvement in the established health system improves the overall outcome of neonates [12,13].
The aim of this study was to assess, from the perspective of key stakeholders comprising leaders in the public health system, the Government of Ethiopia health system's readiness to support health care facilities in the care provided to preterm, LBW, and sick newborns in 3 different regions of Ethiopia. For the purposes of this paper, we define health system readiness as the preparedness of the system, including all components-facilities, staff, and organizational structure-to provide appropriate evidence based care for preterm, LBW, and sick newborns primarily from 28 weeks gestational age and above.

Study design
A team of researchers from Ethiopia and the United States conducted a qualitative assessment in multiple sites across Ethiopia to describe facility readiness to deliver care services to preterm, LBW, and sick newborns, by assessing service availability and key processes.
From December 2017 to February 2018, data collectors conducted key informant interviews with facility leaders in the public health care system, many of whom were also clinical care providers delivering obstetric and newborn care services in study health facilities. Qualitative methods were used to gather rich data to inform how and why the health system needs improvements in serving preterm, LBW, and sick newborns, rather than to establish prevalence numbers. Taking a narrative approach, we used thematic content analysis to answer specific a priori questions about the Ethiopian health system. This report conforms to the Standards for Reporting Qualitative Research [14].
The WHO health systems building blocks provide a framework for understanding and assessing the key components of a health system. Within each block, key domains and the critical factors that influence structure, function, and outcome of that domain are described. This study used a modified version of the health systems building block framework to guide our assessment of the current context of services for preterm, LBW, and sick newborns in Ethiopia [15].

Study sites
The Federal Democratic Republic of Ethiopia is among the most populated countries in Africa, with a population estimate of 107 million for 2018 [16]. One urban and 2 rural regions were selected to allow for assessment and description of different settings and contexts in Ethiopia representing an urban community, a settled agrarian community, and a semi-pastoral agrarian community.
Staff and clients from a subset of health facilities at each tier of the public health care system in each region participated, including specialized hospitals, general hospitals, primary hospitals, health centers, and health posts [17].

Data tools
A new tool was developed for this project using a multistep process. First, based on the WHO framework, the scope of the tool was defined, and questions were drafted. After their initial training, data collectors pretested the instrument at St. Paul's Hospital Millennium Medical College with 5 interviewees similar to the target sample. Adjustments were made for flow, content, terms used, prompts, and instructions.
Using this structured interview guide, facility leaders were asked questions about policies and guidelines, programs, facility preparedness, and referral transfer systems, all with respect to preterm, LBW, and sick newborns (see Additional file 1).

Sampling
We used purposive convenience sampling to recruit from all clinical, nursing, and administrative leaders who were involved in the oversight of units or care programs for pregnancy, labor, and delivery/obstetrics, and gynecological, postnatal, and newborn care services at the study health facilities. We interviewed 37 leaders from all public specialized, general, and primary hospitals and a subset of health centers and health posts located within the 3 study regions (see Table 2).

Saturation
Interview data were reviewed from all 3 study regions periodically during data collection until data saturation [18] was reached, according to the judgment of the core team.

Research team composition, training, and supervision
The research team was composed of core members, data collectors, and data analysts. Experienced data collectors were trained on the study overview, objectives, participant selection, instruments, and interview skills. Supervision was conducted throughout data collection in all 3 study regions using a supervision checklist.

Data collection and management Recruitment
Trained study interviewers identified facility administrators who identified department leaders eligible to participate in the study. Interviewers invited and scheduled potential participants for in-person interviews.

Location
Each interview was conducted in the leader's office at a time convenient for the interviewee. Participants were offered their choice of languages (Amharic or Oromiffa) for the interview.

Recording, transcribing, and translation
Demographic data were recorded on a tablet and interviews were recorded digitally. Recordings were transcribed into the local written language by experienced transcriptionists and were subsequently translated by professional translators. Transcriptions and translations were spot-checked for accuracy by a third team member. Every effort was made to maintain participants' confidentiality both during data collection and during report writing and publication. All audiotapes were destroyed immediately following transcription. No names are attached to any of the data. In the results, quotes are identified only by location (AMH = Amhara, ORO=Oromia, AA = Addis Ababa) where relevant, and by facility level (HP = health post, HC = health center, HOSP = hospital) where relevant to the findings.

Analysis
Data were entered into NVivo version 12 and analyzed using thematic content analysis [19]. Researchers involved in the study design and qualitative coders developed a codebook and applied it to the coding of all transcripts.
All transcripts were open-coded using the final version of the codebook to capture key themes and relevant ideas as identified in the data. Each transcript was coded by 2 separate coders, and disagreements were resolved by the lead analyst who reviewed all discrepancies and discussed them with the second coder as necessary to reconcile coding. Once coding was complete, code reports were produced for each code, cleaned, and prepared for synthesis.
Each code report was synthesized by one team member as follows: initially, text excerpts no longer appearing relevant to the code were grayed out; remaining excerpts were annotated with comments; particularly illustrative quotes were highlighted; and comments were summarized into a table of theme domains and subdomains with associated quotes for each code report.

Description of sample
All respondents were from Addis Ababa, Amhara Region, and Oromia Region. The characteristics of study regions are shown in Table 1.
Clinical leaders were proportionally divided across all 3 regions; 50% were 26 to 30 years of age; and all were employed as midwives, nurses, health officers, heath extension workers (HEWs), or neonatologists (Table 2).

Health service delivery Infrastructure
Across the public health system, adequate and reliable space, power, and water functionality were problematic. Table 3 displays the issues linked by facility levels.

Service readiness
Participants were asked how ready their local health system was to deliver (1) routine care to newborns and (2) care to preterm, LBW, and sick newborns-across labor and delivery, NICU, kangaroo mother care (KMC), and inpatient wards. Only hospitals and advanced health centers have specialized care units, but all leaders were asked about readiness for newborn care and preterm, LBW, and sick newborn care. A few leaders reported being very equipped to provide care, but most cited problems in one or more areas that inhibited effective provision of care. Table 4 shows the challenges related to readiness in each of the facility units.

Health workforce and human resources
Human resource issues (training, staffing, and retention) were critical to being able to properly care for preterm, LBW, and sick newborns.

Provider training and skills
Clinical leaders saw staff training as mandatory for quality service provision and specifically recommended training in newborn care, including refresher training that did not currently exist: "There should be constant providers and the number of providers need to be increased; one provider is needed for one baby if they are critically sick; if possible one nurse should be assigned for one baby. To do so, the number of health providers matters" (AA-HOSP). Training on newborn care basics was not common for staff in general. Staff preservice training in the preterm, LBW, and sick newborn program was inadequate for many, if not most, providers. Labor and delivery training was limited and training applicable to preterm, LBW, and sick newborn care was primarily related to breastfeeding and KMC. Previously, nongovernmental organizations had provided some training in care for preterm, LBW, and sick newborns, but that was no longer happening.
Lack of training was identified as one factor leading to unnecessary referrals. Inadequate numbers of trained providers-especially senior staff, skilled professionals, and NICU staff-were also seen as limitations to system readiness. The difficulty of getting a newborn with a complication to an appropriate specialist led to loss of life. Furthermore, high turnover of trained professionals was an ongoing problem.

Communication
Functionality also depended upon effective communication between providers in the same facility and in other facilities. Participants reported that meetings were the Participants recognized the importance of communication for effective care of preterm, LBW, and sick newborns and called for improvements both within units at facilities and between levels of facilities: "To improve these challenges, I think we need to communicate-I mean, all of us in our respective departments/units. The doctors with the doctors working there, the interns with the interns working there have to closely communicate. The nurses and midwives also have to communicate among themselves on service delivery and on solving the problems" (ORO-HOSP).
Communication followed a hierarchical path when information or expectations were being transmitted from the top of the system down through the various facility levels, as one participant described: "Orders trickle down from the authorities at the zone level to the ones at the woreda (district) level, then to the health directory, then to the health center, then finally to the health posts. These orders are usually transmitted orally but there are cases when we file them as written forms. We will be given instructions, updates on policies and regulation so we can apply them to our model. We are also given guidance on the best route to implement the reforms, then we inform the community on the steps to be taken" (ORO-HP). It was not clear how information, requests, and expectations were conveyed from the bottom up in this system.

Health extension workers
HEWs were seen as the key link between mothers in the local community and the entire health system through initial identification during pregnancy and postpartum follow-up. In the health system design, Health Development Army volunteers were supposed to identify pregnant women or recent home births and pass that information to the HEWs. HEWs then performed the critical role of home visits and monitoring mothers and babies in the community. They sometimes managed home delivery issues and follow-up care for preterm, LBW, and sick newborns: "Since the primary responsibility of the health extension worker is within the community in rural [areas], they are accountable to follow newborn and maternal health conditions. Them being there will support us, because they will take the responsibility to ensure pregnant mothers initiate follow-up and come to the health center for childbirth. Then we will follow up, calculate the EDD [estimated date of delivery], and identify those who will come for follow-up visits or not. Further, the health center will decide which newborn case can be treated here and transfer those which need a critical follow-up. That is how we do things" (AMH-HC).
HEWs played a critical role in referring preterm, LBW, and sick newborns to higher levels of care, especially in rural areas and especially following home births. Several participants spoke of the need for leadership to recognize this critical role and prioritize it. In some cases, HEWs accompanied patients to higher-level care.  Mothers not able to clean up after birth "It is difficult for a mother to go home covered in blood after birth. .. For example, if a mother gives birth here on dusty space, then it is no different from giving birth at home" (AMH-HC) Shortages for 1 to 2 weeks "We may not get water for 1 or 2 weeks. To eat our food, we have to buy packed water. Even we do not get to wash our hands. We prepared large water container, so we fetch from that. .. We have to carry from the ground [floor] to the second floor" (AA-HOSP) Periodic interruptions due to an aged building Parents restricted from visiting preterm, LBW, and sick newborns in NICU if there was no water to wash visitor gowns When women self-referred to higher care, it was seen as a failure of the HEW identification and tracking system. Participants made several recommendations to improve the HEW role and functionality in the care of preterm, LBW, and sick newborns: Conduct initial and ongoing training and capacity building; this is necessary to increase skills if HEWs take on expanded, more sophisticated roles; Improve referral system feedback to strengthen the relationship between HEWs and higher-level facilities; Increase supervision and support for HEWs; Increase governmental support for the role of HEWs and commitment to sustained improvements; and Demonstrate leadership respect for HEWs so that others within the larger system adopt a respectful attitude toward HEWs.
Participants also noted that the performance of HEWs could be improved only within the limits of the current system's infrastructure; problems related to the functionality of water, power, and roads would limit even the best trained and supported HEWs.

Health management information systems and data monitoring and evaluation
Data were being collected at multiple levels throughout the health system, but not consistently within levelsnor even within facilities in some cases. All 3 facility levels-health post, health center, and hospital-reported at least one example of using data for performance monitoring, quality improvement, or health trends. Participants were aware that documentation was important, not only for individual patient tracking but also for health trend tracking, such as observing preterm birth rates, for larger system planning.

Information production
Data collection practices were problematic on many levels. Primary documentation seemed to be associated with referral, although this may be due to the focus of the interview instrument on referrals. The data collection system was variable, and even erratic, depending on the individual facility and the level of facility. A participant from a health post commented, "Another thing is since we don't have a standard referral form, some information may be left out, which is dangerous" (ORO-HP). Data were recorded on patient cards, log books, booklets, referral forms, and plain letter paper. It was unclear how data were tabulated, collected, and maintained. Health posts played a critical role in gathering data and documenting patient visits. At the kebele (ward) level, data were recorded in 2 places: the booklet of diagnosis guidelines, and the newborn registration log books (for future follow-up by HEWs for antenatal care). Health post staff were aware that proper documentation informed quality supervision and performance monitoring.

Information use
Ongoing performance monitoring appeared to be happening especially in higher-level facilities based on observation, but not necessarily monitoring data. Several respondents supported the idea of performance monitoring and called for more: "Such things should not be periodic, it should be continuous and the top leaders should design a control system for lower-level implementers. So it is good if it has monitoring and evaluation" (AA-HC). They recognized that performance monitoring could reveal training effectiveness. One urban health center spoke about a strong performance monitoring link between the hospital and health center, and with the FMOH.
Access to essential medicines, supplies, and equipment Equipment functionality and distribution Problems with dysfunctional equipment and ineffective equipment distribution systems were widespread. Not only was equipment not being purchased, but when purchases were made without consulting or involving providers, they risked being unusable. In the past, nongovernmental organizations provided support to borrow equipment, but this was not a sustainable solution.
Equipment, materials, and medications were unavailable, expired, or broken at all levels of the system. For the purposes of this report, ambulances are classified as equipment because respondents referred to them as such. Few facility leaders said they were well equipped. Problems at facility levels were as follows: Health posts: No available weight scale, timer, breath control, blood pressure monitor, radiant heater, phone network, or ambulance, and incomplete medication supplies. Some equipment was provided by nongovernmental organizations but not by the FMOH; some equipment was broken or very old. Health centers: No available oxygen supply, radiant heater, incubator, ambulance, phone network, lab reagents, soap, water, or towels; medications were expired or limited. Almost every health center respondent said that ambulances were one or more of the following: unavailable or delayed, lacked oxygen, relied on motorcycles, or experienced big delays; only one health center said ambulances were not a problem. Hospitals: No available infusion syringe pump, portable X-ray, oxygen tubes, pulse oximeter, bag valve mask resuscitator, mechanical ventilator, continuous positive airway pressure, incubator, phone, diapers, baby clothes, or personal protective equipment for staff exposed to radiation and infection. Newborns were exposed to infection by sharing beds, incubators, and phototherapy machines. There were shortages of ambulances and NICU supplies, such as infusion syringe pumps, nasal prongs for preterm, bag valve mask resuscitators, vitamin K, and glucometers.

Health financing
Respondents were nearly unanimous that there was no sustained financial support for preterm, LBW, and sick newborn care in facilities. Hospital leaders talked about being able to include it in the general budget, but they asserted that there was no separate funding support for programs for preterm, LBW, and sick newborns. One health center representative observed: "I think there needs to be a properly budgeted program which rigorously works to increase awareness in the society. Simultaneously, the health center needs to support facilities and even ambulances so that it can provide the level of health care service promised to the community" (ORO-HC). Funding needs at health centers and hospitals included: specialist salaries, printing of guidelines and policies, basic sick baby supplies (e.g., milk and clothing),  Leadership and governance Guidelines, policies, and protocols Participants were asked a series of questions about guidelines, clinical standards, protocols, or training modules relevant to prevention, management, or care of preterm, LBW, and sick newborns-specific to the level of facility they supervised. Responses were analyzed to assess how the facility leadership prioritized use of guidelines and to what degree guidelines were available and being used. One big challenge was that leadership across facilities interpreted the concepts of guidelines, policies, and manuals differently. At health posts, references to guidelines primarily included training manuals and chart booklets, while at health centers and hospitals, guidelines might include published manuals, guidelines, and protocols. Availability of guidelines varied widely across facility levels, within the same level but between facilities, and even within facilities. There was no consistency anywhere in the system regarding guidelines for care of preterm, LBW, and sick newborns. Perhaps the most widely available guidelines were related to referrals, some of which were very simplified at the health post level and accessible in chart booklets. This is not to say that other guidelines did not exist, but they were not universally available. Most importantly, if newborn guidelines were available, they were not specific to preterm, LBW, and sick newborns. Use of guidelines depended primarily upon availability. When guidelines were available, they were seen as an important resource: "We provide the guideline for the assigned provider at the unit and we make sure the guideline stays in the unit; we even have a locker for that so to make sure anybody else assigned at that unit can access the guideline. Because we can't work without the guideline, the guideline is required critically. The guideline and the log book are inseparable documents" (AMH-HP). Guidelines were used for assessment and referral, resuscitation, parent access to baby, complicated case management, and treatment. One health center representative spoke about posting key points from the guidelines in the medical compound for easy reference. Some spoke about using outdated guidelines because they saw that as better than no guidelines at all. Nearly all those who were aware of, and knowledgeable about, guidelines and policies found them clear and understandable.
For those who did not use or adhere to the guidelines, across all facility levels, reasons ranged from poor dissemination to outdatedness. Respondents recognized the need for ongoing guideline dissemination: "These materials-according to the system, it could be from the ministry of health to the health bureau to the sub-cities and woredas-should flow to us timely. And the updates should be properly introduced and training the providers is needed to introduce the providers with new materials so that they can primarily focus on the programs" (AA-HC). When asked about what might facilitate expanded usage, participants had several key suggestions. Challenges and facilitators for guideline use are in Table 5.

Supervision and feedback
Staff supervision and feedback were valued by participants, but responses suggested the feedback needs to be systematized and bolstered at all levels across and within facilities. Feedback was primarily through patient referrals. Feedback at catchment meetings was recognized as an opportunity to provide guidance on areas for improvement as well as recognition of accomplishments from the hospital to the health center. Similarly, feedback between health posts and Health Development Army leaders offered opportunities to reduce the number of out-of-facility births. Supportive supervision and mentorship were valued but appeared to come primarily from external sources rather than from direct supervisors or leadership. More stringent supervision and mentoring were needed universally, and especially in NICUs, where highfunctioning staff were critical. At the health post level, participants requested that district-level supervisors support health posts when supporting programs from tangential sectors such as agriculture, and also, that they provide follow-up supervision and feedback after trainings.

Leadership priorities
Participants were in nearly universal agreement that maternal and child health was a focus priority for the FMOH, citing evidence that emphasis was placed on NICU training, referral, and transport issues. The link between hospitals and health centers helped to maintain the maternal and child health focus. For maternal and child health to be considered a priority focus, however, the FMOH would need to secure additional funding support (including funds to meet newborns' needs); establish an organized program for preterm, LBW, and sick newborns; and ensure that leadership would drive the program forward. Although respondents noted that treatment and management of preterm, LBW, and sick newborns were considered FMOH priorities, a great gulf remained between stated priorities and availability of materials, recommendations, and training to implement programs. As one respondent stated, "The government has prioritized maternal and child services by recognizing preterm or LBW newborns' need for early treatment and management, but we are not providing all necessary services here for newborns. .. For the health center to improve service, there should be a protocol; we will get confused to treat or manage such cases in current situations. We will be more confident when necessary materials are adequately available. .. In addition, training should be provided to professionals because knowledge is important for service. Knowledge leads to skill" (AMH-HC).

Community engagement
Community engagement is not one of the WHO health systems building blocks, but it is vital to providing ongoing care. Participants reported that community engagement and awareness were accomplished through key individuals in the communities: Health Development Army volunteers, religious leaders, traditional birth attendants, and midwives. As one leader suggested, "We should select some influential family members and community members and make them part of trainings to disseminate the information to the community through the community members" (AMH-HP). Additionally, providers recognized that responding to extended families in hospital settings and relying on women who had successfully delivered in facilities were informal ways to spread goodwill between the health system and community. As one health center participant explained, "The most difficult thing for us is that women seem to believe and rely on what they talk about amongst themselves rather than what we teach them. For them we are regarded as outsiders, so we try to make those women from the community who have had health care services speak out on our behalf, and it has proved to be successful" (ORO-HC). Engagement was seen as particularly difficult when the community was mobile. Topics recommended for community awareness included home birth risks, importance of medical care for newborns, and family planning. Participants also acknowledged that leaders must allocate budgets to create awareness. Some staff at facilities described efforts to engage the community by showing respect for cultural traditions, like coffee ceremonies and porridge making, and by observing Neonatal Day as a holiday. These efforts to draw mothers and families into the health system were up against the strong pull of traditions that limit access to facilities, including prohibiting mothers from leaving home for 40 days following birth, and concerns about babies or mothers dying in a facility far from home.
Communication with families was recognized as vital to health system readiness to serve families of preterm, LBW, and sick newborns. Participants reported that providers' communication with families in clinical settings varied widely depending on whether responding to their questions or providing updates or information. Communication may have varied widely also by facility type or site. Table 6 shows how respondents saw communication playing out between providers, facilities, and families.

Programs for preterm, LBW, and sick newborns: existence and components
Participants were asked about the existence and components of programs for preterm, LBW, and sick newborns. Most respondents said their facility had no program of which they were aware. Health post leaders mentioned programs for referral and postnatal education, and health center personnel mentioned equipment (heater, ambulance, and booklets), supervision, the referral process, and previously existing programs. Hospital leaders discussed complex medical programs they currently had. All respondents recognized the importance of focused programs for preterm, LBW, and sick newborns and requested support. As one person said, "First the programs should be initiated; I don't think the programs are beyond paper" (AMH-HC). Another called for a coordinated effort: "First, there should be trained personnel, then materials that help to provide services for newborns should be adequately available including medications. .. moreover, this program should have a focal person who works on maternal and child health. .. and a strengthened follow-up system must be placed. I think different stakeholders including woreda, zone health departments, and partner organizations must be involved to provide regular support" (AMH-HP).

FMOH support for preterm, LBW, and sick newborn programs
The FMOH has clearly stated that maternal and child health is a public health priority, but that has not positively affected staff strengthening, funding, or materials provision that would support initiating or continuing programs for preterm, LBW, and sick newborns. Respondents unanimously agreed that no funds were available from the FMOH for initiating any kind of program for preterm, LBW, and sick newborns at any facilities. Similarly, no respondents were aware of any staff-related support for preterm, LBW, and sick newborn program activities, with the exception of limited training opportunities. Responses were nearly unanimous that there was no support for materials or training guides. A few people referred to initial training or booklets they had received, but they were primarily for midwifery training on newborn care rather than care for preterm, LBW, and sick newborns.
The lack of sustained support for preterm, LBW, and sick newborn programs created a reliance on nongovernmental organizations to fill that gap. The FMOH has not consistently supplied equipment or guides to support ongoing preterm, LBW, and sick newborn programs. Heaters have been supplied to some facilities, but there was also a call for newborn clothing and medications. One person referred to the unsustainable practice of borrowing equipment, and others described equipment that may have gone unused due to lack of training. Table 7 presents key findings and recommendations.

Discussion
This qualitative study assessed Ethiopia's health system readiness to support public health facility care of preterm, LBW, and sick newborns. Each of the WHO health system building blocks were considered, and findings revealed that public hospitals, health centers, and health posts faced significant facility and system challenges in all 7 components, making optimum care for preterm, LBW, and sick newborns difficult. Unlike other studies looking at multiple countries, in this study we gathered in-depth information across 3 economically and geographically diverse regions of one country. The findings from this study provide evidence to inform action steps to address complex system-level challenges.
Over the past decade, WHO and others have given much attention to health systems strengthening [7,15], and the Ethiopian government has reaffirmed commitments and consolidated gains [22]. We relied on facility leadership-the immediate recipients of health system inputs to facilities-as our main data source, building on findings from other studies that health care professionals are key information sources to monitor progress in health systems [23,24].

Health service delivery infrastructure and readiness
Ethiopia has achieved significant expansion of all levels of public health facilities in the last 2 decades [17]. Facilities are reported to be well equipped to provide care, but most cited problems in one or more critical areas that inhibited effective provision of care. Reliable power Table 6 Communication between families and facility staff: domains and subdomains

Domains Subdomains
Updating mother/father, sometimes family Updating done frequently but not during emergency procedures Recognition that good updating was part of compassionate and respectful care Communicating bad news to parents could be very challenging Responsibility for updating parents varied by specialist, general practitioner, or nurse, depending on who was attending to the baby Parent/family questions policy Parents were encouraged to ask about newborns anytime; questions from extended family were also answered on demand Where policies existed about communicating baby status to parents, they were usually followed; some facilities had no policy and water have remained major problems, especially in lower-level facilities [25]. Space in health facilities for newborns in need of special care, including space for mothers to stay with their babies, was either absent, inadequate, or lacked privacy. Similar problems were reported in other African and Asian countries [9].

Workforce/human resources
Health workforce shortages were found to be a critical obstacle to delivering quality care for preterm, LBW, and sick newborns. As found in other reports [8,9], nationwide shortages of adequate and well-trained health professionals of almost every category were found at all levels of health facilities. The problem was worse in rural and remote regions of the country and at the lowest levels of the health system. The number of neonatologists, pediatricians, and neonatal nurse specialists in the country were few and concentrated in tertiary centers [26]. This problem offered 2 equally important solutions: produce more graduates, and continuously strengthen the capacity of existing staff working in remote and lower-level health facilities. Trainings and the national programs supporting them were integrated with general newborn care and not specifically focused on preterm, LBW, and sick newborn care [27]. Findings also emphasized the important role of HEWs in • M&E data were collected at various levels within the health system inconsistently and irregularly • Available data were not always used for performance monitoring and quality improvement • Improve the consistency and quality of data collection and use of data at all levels of health system Access to essential medicines, supplies, equipment • Shortages of medical supplies, equipment and essential medications were widespread in health facilities at all levels • They were unavailable, broken, or inappropriate for use • Prioritize procurement and maintenance of critical supplies and equipment

Financing
• Funding allocated for system readiness to care for preterm, LBW, and sick newborns was needed in nearly all facilities at all levels • Specific funding needs included specialist salaries, printing of guidelines and policies, basic sick baby supplies (e.g., milk and clothing), transportation for low-income patients (other than transfer ambulances), food for patients and parents, phone service for providers.
• Fund a program specifically to support preterm, LBW, and sick newborn care at the facility level, which may help to alleviate challenges and ultimately improve the available care Leadership and governance • Gap in the availability of guidelines and protocols specifically targeting preterm, LBW, and sick newborns • Staff supervision and feedback were valued by participants, but responses suggested they need to be systematized and bolstered at all levels across and within facilities • Supportive supervision and mentorship were valued but appeared to come primarily from external sources rather than from direct facility leadership • Develop evidence-based, up-to-date guidelines and protocols specific to the care of preterm, LBW, and sick newborns, and communicate them across the system tier with appropriate supervision • Develop more stringent supervision and mentoring, especially in NICUs where high-functioning staff are critical Community engagement • Key facilitators and information disseminators/influencers identified in the study were the Health Development Army, community and religious leaders, and mothers and families who had positive experiences or outcomes of care • Showing respect for the community's traditions was recognized as an effort to positively change the perception of the community • Improve awareness through health education, peer modeling, and dissemination of good experiences Abbreviations: LBW low birth weight, M&E monitoring and evaluation, NICU neonatal intensive care unit, WHO World Health Organization identifying and referring preterm, LBW, and sick newborns, and the need for recognition and continuous support of HEWs.

Supplies, equipment, and essential medications
Shortages of medical supplies, equipment, and essential medications were a widespread problem in health facilities at all levels. They were often unavailable, broken, or inappropriate for use. In another report, fewer than half of all facilities had most supplies and equipment needed for preterm, LBW, and sick newborns, and these facilities were primarily specialized hospitals rather than general and primary hospitals [28]. In another report, 35% of Ethiopian health facilities experienced stock-outs of essential medicines, and approximately 8% of medicines in stock had expired [29].

Health financing
This study showed that specific funding allocated for system readiness to care for preterm, LBW, and sick newborns was needed in nearly all facilities at all levels.
For families, the birth of preterm, LBW, and sick newborns could be financially catastrophic. Shifting from a reliance on out-of-pocket payment to prepayment and risk pooling is a critical part of the health financing transition that most countries go through as they get richer. The Ethiopia health care financing strategy aims to replace user fees with funding from the central government. This is expected to boost utilization of services and consequently increase demand on resources [30]. The health care financing strategy targets reduction of out-of-pocket expenditures for direct clinical care services costs, however, and does not include indirect costs for supplies in the form of clothing and formula milk for feeding (when indicated) and emergency transportation costs-which is important for saving lives in cases of emergency by facilitating travel to health facilities.

Leadership and governance
The availability of guidelines, policies, and protocols, and the practice of supervision and feedback were used as measures of effective leadership and governance. There was a notable gap in the availability of guidelines and protocols specifically targeting preterm, LBW, and sick newborns. From a recent analysis of health system bottlenecks, policy documents in circulation among senior officials were not disseminated to lower-level managers who did not always adhere to guidelines for special newborn care [9]. This study, too, found that in Ethiopia, narrow dissemination and outdated materials contributed to poor adherence to national protocols and guidelines. Another study in Ethiopia reported that hospital activities were poorly supported with a lack of updated or appropriate guidelines and protocols [31].
Staff supervision and feedback were valued by participants, but responses suggested they need to be systematized and bolstered at all levels across and within facilities. Most feedback was related to referrals, and participants spoke positively about effective systems between facilities, especially case-specific feedback. Functional referral systems are crucial for the referral of all patients who need care that cannot be provided in the health facility. An essential component of quality of care for preterm, LBW, and sick newborns is proper handover to the receiving facility and feedback to the sending facility when completing a referral. This is a global standard for improving quality of service for preterm, LBW, and sick newborns [32].
Supportive supervision and mentorship were valued but appeared to come primarily from external sources rather than from direct facility leadership. More stringent supervision and mentoring were suggested universally and especially in NICUs where high-functioning staff are critical. In a study in India, nursing staff felt that there were not enough proficient providers to mentor the nurses to safely perform skills, and they identified lack of mentorship as an obstacle for improving service delivery for preterm, LBW, and sick newborns [33]. Ongoing on-site training, mentoring, and supervision were also cited as absolutely necessary to maintaining the knowledge and motivation of mid-level health care providers and reducing neonatal mortality [34,35].
Ethiopia's National Neonatal and Child Survival Strategy aims to reduce the neonatal mortality rate from 28 deaths per 1000 live births to 11 deaths per 1000 live births [6]. Although some interventions like KMC, which were defined in this strategy, may have benefited preterm, LBW, and sick newborns, existing programs were not specifically focused on supporting the care of such newborns in all levels of health facilities.

Data monitoring and evaluation/health management information systems
This study showed inconsistent and irregular monitoring and evaluation data collection at different levels within the health system. Available data were not always used for performance monitoring and quality improvement. The use of health management information systems is important in developing health care services, but many LMICs struggle with data quality problems resulting in incomplete, inaccurate, and untimely information, which is not useful for health decision making [36]. In Ethiopia where the health management information system is the major source of health information [37], the overall data quality was found to be below international standards [38,39]. These findings also showed the need for improvements in data quality and usefulness.

Community engagement
This study revealed that community engagement was considered an important component of serving families with preterm, LBW, and sick newborns. The key facilitators, information disseminators, and influencers identified in the study were the Health Development Army, community and religious leaders, and mothers and families who had had positive experiences or outcomes of care. Showing respect for the community's traditions was recognized as an effort to positively change the perception of the community. Community engagement in a mobile society was found to be a challenge. The findings were in line with the FMOH strategy to reduce overall neonatal mortality [6]. While engaging community in the care of preterm, LBW, and sick newborns is reported as a major obstacle in many countries [9], a recent study in Malawi showed a shift in perception of community members through health counseling, peer modeling, and personal success with KMC [40].

Limitations
The input from facility leaders reflects their experiences and expectations, and possibly their optimism or frustration. Expanding the sources of information about the functioning of the health system was beyond our resources. Additional evidence is available from existing hospital data, central reporting systems, and central government representatives. The proportion of providers in facility leadership roles who participated in the interviews may not be representative of all clinicians and providers in the facilities. Information obtained from interviews was not verified by observation.

Conclusion
The Ethiopian health system has challenges in terms of its readiness to support health facilities to provide quality care for preterm, LBW, and sick newborns. Challenges were reported in all 7 WHO health system building blocks-leadership and governance, health financing, health workforce, essential medical products and technologies, health service delivery, health management information system, and community partnership. Many of the readiness challenges across the health system seem to stem from the inadequate emphasis given to the special care required for preterm, LBW, and sick newborns.

Recommendations
Based on the findings of this study, we propose the following recommendations and suggested actions (see Table 8).