Even though there has been substantial decline in deaths of infants and young children in the last two decades, the burden of disease and mortality remains high(1). In 2019, almost half, two and half million of deaths in young children occurred during the first month of life, (neonatal period; 0–28 days), with the vast majority (> 90%) occurring in low- and middle-income countries (LMICs)(2). The most common causes of these neonatal deaths are prematurity, intrapartum complications, and neonatal sepsis.
In Every Newborn Action Plan (ENAP) of 2014 and Sustainable Development Goal 3.2, countries made a commitment to reduce the neonatal mortality rate to at least as low as 12 deaths per 1000 live births by 2030(3, 4). For this to be achieved, countries need to scale up coverage and implementation of evidence-based high impact interventions (Kangaroo mother care, treatment of small for gestational age and sick neonates)(5). Quality data is required to keep track of this coverage and implementation. There are challenges with neonatal care data in LMICs, for example data is not readily available from routine health information systems [mainly collected in health surveys] and where available, the quality is poor due to lack of standardized registers across various hospitals and non-linkage of datasets within and across hospitals(6).
A 2015 study by Aluvaala and others estimated the main causes of admission in 22 Kenyan County hospitals (Formerly District hospitals) to be birth asphyxia (36%), prematurity/low birth weight (32%) and neonatal sepsis (19%)(7). There have been improvements in newborn health with recent estimates across 16 hospitals in Kenya showing the main causes of newborn admission as intrapartum-related complications (30%), respiratory distress syndrome (18%), neonatal sepsis (15%), jaundice (12%) and uncomplicated low birth weight (LBW)/prematurity (5%) (8). In both of these studies, comorbidities were common, and the newborns were on multiple interventions as part of newborn care within the newborn admission ward/newborn unit.
Care for neonates in Kenya is guided by policies made at the national level (National government) and service provision by County governments (public sector) and other private and non-governmental health facilities. Despite infrastructural and capacity investment for neonatal care in Kenya over the last few years, neonatal deaths accounted for two thirds of children admitted in the 16 hospitals with more than half of these deaths occurring within 24 hours of admission(8). There is a need to continuously undertake research on how to enhance quality of care during this neonatal period. One of the key undertakings by policy makers is to develop clinical practice guidelines and data collection tools and registers that guide neonatal service provision. Kenya has adopted the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach to the development of pediatric clinical guidelines. This approach is heavily reliant on comparable outcomes across studies to allow for proper evidence synthesis. During the last revision of the Kenyan pediatric protocols (which includes neonates), in the review of evidence on neonatal care, heterogeneity of outcome data on neonatal sepsis was noted as a key limitation to conducting a meta-analysis, leading to policy makers depending on low to moderate quality of evidence for the recommendation on umbilical cord care(9). There was also scarcity of routine health system data on neonatal care. To enhance research synthesis for clinical guidelines and enable setting up of pragmatic trials, outcomes that are reported in neonatal care and research need to be comparable(10).
One of the ways to achieve this comparability of outcomes in neonatal care is to ensure that we have standardized outcomes to guide neonatal research and clinical care in LMICs.
What are COS?
A core outcome set (COS) is an agreed-on minimum standardized outcome set(s) that should be measured and reported in all research in a given health area (11). A COS consists of a core domain set (this defines what domains should be measured in a trial) and core measurement tools (defining how the outcomes should be measured). Core outcomes are important outcomes agreed on by key stakeholders (including patients or public participants) using robust consensus building methods. Use of COS could ensure all future research and routine systems collect not only clinically important outcomes but also help to ensure inclusion of outcomes that are important to patients/caregivers and policy formulators. This could reduce research waste, by facilitating meta-analysis, minimizing reporting bias and thereby enhancing research translation and use (12). COS are also becoming useful in routine data collection systems, which enhances designing quality of neonatal care initiatives through quality audits and feedback (13, 14). It is worth noting that since a COS is a minimum set, it does not limit researchers and clinicians reporting other outcomes which they may deem relevant. COS use is becoming increasingly recommended by research funders, journal editors and clinical guideline developers in high income countries (HICs) with the potential to reduce research waste and improve the quality of research reporting (15).
COS development
Globally, COS have been developed for various conditions or diseases and continue to be developed (16, 17). There has been an increase in the inclusion of LMIC stakeholders in COS development and use. However, most COS work has been undertaken in HICs with only 20% of the 370 COS registered in the COMET Initiative database having stakeholders from LMICs. Only four COS have been initiated from LMICs (18). LMICs often have different burden of disease, healthcare systems, resources, and research infrastructure. This could in effect mean that the COS identified in a HIC setting for a given disease area may not necessarily be applicable in LMICs.
Some of the reasons cited for low initiation of COS and inclusion of stakeholders from LMICs are lack of knowledge on COS availability and utility, health system issues like lack of tools or means to measure the agreed-on outcomes, inadequate funding to fully engage stakeholders (19). It is also postulated that, literacy levels, use of technology and lack of formal patient organizations or associations could reduce involvement of patients or lay public though optimal patient involvement strategies are still being developed(20).
Transferability of COS developed in HICs to LMICs
It is increasingly realized that to make research more relevant to practice, contextualization is of great importance. Encouraging and normalizing the development and use of relevant COS by researchers, clinicians, funders, and policy makers is a key priority to make the COS more globally applicable. As noted above, LMICs stakeholder inclusion has expanded over the last few years of COS development to include more patients, care givers and lay public. There has been a cross linking of COS for research with clinical practice with the aim of using the COS in clinical audit and feedback for improvement of care(21). There is however paucity of evidence as to whether the methods for COS development that have worked in HICs can be used in LMICs and whether a COS developed in a HIC can be used as is in an LMIC or it would require adaptation.
Only one COS has been led from an African setting. This is a COS on congenital abnormalities as part of setting up a congenital abnormalities’ registry in Rwanda (22). However, as of end of 2019, of the 370 COS registered in the COMET database, 22 had stakeholders from African countries though they were mainly drawn from South Africa (18). No COS has been initiated in Kenya.
Why neonatal COS?
With a current nationally estimated 20 deaths per 1000 livebirths in the neonatal period, Kenya has a long way to go if the target of at least as low as 12 deaths per 1000 live births is to be achieved(25). Even though there is an improvement in facility-based deliveries and by extension care of newborns and neonates, neonatal deaths still account for two thirds of pediatric admissions deaths in Kenya(8). Improvement of neonatal care needs to be enhanced and multipronged. For this to happen, the right quality data is needed to ensure that there are targeted interventions. One of the challenges faced by policy makers and clinicians getting the right quality data is the lack of standardized outcomes being reported not just for routine newborn care but also for neonatal research. One of the ways to have these standardized outcomes is to have a minimum set of outcomes being reported in routine practice and research. This minimum set needs to be agreeable to the key stakeholders in the local context for implementation to be efficient.
Neonatal Care COS
There exists a COS for neonatal care that was developed in 2019 by Webbe and others(23). The study set out to define a core outcome set (COS) for research involving infants receiving neonatal care in a high resource setting. The process included a three-round e-Delphi survey[2] with 173 participants and a face-to-face consensus meeting of 16 people, to confirm the final COS based on the survey results. The participants were former patients cared for in a neonatal unit, and parents of neonatal patients, doctors, nurses, and researchers.
While some stakeholders from LMICs were included in the development process, their inclusion was based on their prior experience of neonatal care or research in a high resource setting neonatal unit. Twelve outcomes were included in the final COS: survival, sepsis, necrotizing enterocolitis, brain injury on imaging, general gross motor ability, general cognitive ability, quality of life, adverse events, visual impairment/blindness, hearing impairment/deafness [for all neonates], retinopathy of prematurity and chronic lung disease/bronchopulmonary dysplasia [for preterm babies only]. The study did not provide recommendations for how these outcomes should be measured.
Given that this COS was developed from a high resource perspective, it is unclear whether, the COS may be directly transferable into LMIC settings due to several reasons:
The differing resource availability for running neonatal care units may mean that the quality of care provided and therefore positive outcomes are more likely in HICs compared to LMICs, and therefore the outcomes important to parents in HICs settings might differ to those of parents in LMICs.
Some of the outcomes suggested for the COS may already be being measured routinely which may not be the case in LMICs (for example imaging in neonates)(24);
The differing epidemiology of neonatal diseases, for example in LMICs there are a number of underlying comorbidities, may mean neonates would be given several interventions at the same time and this may lead to different outcomes in this setting.
Due to the availability of systems to capture information on the neonates, enabling follow up of patients, the authors of the HIC COS study were able to obtain opinions of former neonatal unit patients. This follow-up information is often lacking in LMICs. Indeed, HIC systems are more able to capture long term outcomes compared to LMICs where short term outcomes may be more likely to be identified.
In the development of this COS, policy makers and hospital administrators were not included as participating stakeholders. It is possible that, when this group of stakeholders are involved, the set of outcomes that are useful to them might be different. Policy makers determine indicators being collected by the routine health system and determine resource allocation which affects data collection and availability (adequacy of health workforce, availability of standard registers tools etc).
There is a need to demonstrate the feasibility (or not) of either adopting or adapting a COS for neonatal care and research developed in a HIC setting for use in an LMIC setting. This work is important as it will help describe a methodology that can be applied to other COS that have had low participation from LMIC stakeholders but are important in LMIC settings.
In this study, we will engage key stakeholders, parents/caregivers of neonates, clinicians managing neonates in newborn units, researchers undertaking neonatal research and policy makers responsible for newborn and child-health policies in Kenya to understand if neonatal care outcomes identified for a HIC setting should be measured (adopted) or would require adaptation in a low resource setting. Additionally, we will assess the feasibility of the routine health information system to collect data for the neonatal care COS.
Rationale
Even though a COS for neonatal care exists, it was developed for use in HIC newborn care(23). Further, stakeholders that were involved in the neonatal care COS for HICs were mainly LMIC researchers and clinicians who had experience in neonatal care or research in HIC. There were no policy makers or hospital managers involved in that process.
It is important to document the process of adapting an existing neonatal care COS and describe the contextual issues that need to be addressed for the adaptation to be successful.
We propose to use qualitative inquiries and a face-to-face consensus building process in determining contextual issues that are important for the various stakeholders in Kenya to ensure that all important outcomes are measured and reported in neonatal care research and practice. Once these outcomes are agreed on, it is useful that we assess whether, a routine health system is already collecting the data or is capable of collection this data. Since this work is constrained within a PhD timeline, it may not be feasible to use the routine Kenya Health Information System. We will therefore use the Clinical Information Network (CIN) to assess this. CIN brings together 22 county hospitals that form the first line of referral in Kenya. It promotes the generation and use of high-quality routine information on hospital admissions to paediatric and neonatal wards as part of a learning health systems(26). CIN uses these data to promote and track adherence to guidelines provides 3-monthly audit and feedback reports on key quality of care indicators as a means of quality improvement(27).
Aim
The overall objective of this work will be to determine outcomes that are useful for neonatal care in Kenya to inform whether an existing COS should be adopted or adapted and assess feasibility of a routine data collection system to collect the COS data.
Specific Objectives
-
To identify outcomes being reported in neonatal care research conducted in sub-Saharan Africa through a rapid review of available literature.
-
To explore the important outcomes for key stakeholders (parents/caregivers of children who are under neonatal care, clinicians, researchers, and policy makers) in Kenya.
-
To undertake a consensus building process on the important outcomes for neonatal care in Kenya.
-
To assess the feasibility of collecting the proposed COS data using a routine data reporting system (CIN).