Understanding what has worked well and what needs to work better to ensure newborns survive and thrive in humanitarian settings requires stakeholder commitment, reliable data, and sustained funding for both health service delivery and implementation research that includes conflict-affected, forcibly displaced, marginalized and stateless populations often missed by national health systems, surveys and surveillance systems. Although residents of refugee camps constitute a relatively small proportion of conflict-affected, forcibly displaced, marginalized and stateless populations that go uncounted in population-based surveys and vital registration systems, reporting and analyzing neonatal mortality burden and trends among this population is a critical step in strengthening data use and accountability for neonatal health in humanitarian settings. The UNHCR Health Information System provides insights into neonatal mortality burden and trends among refugee camp residents, as well as issues to consider in design and use of routine health information systems to monitor neonatal health in sub-national populations.
Although NMRs in stable refugee camps are often lower than surrounding host populations, the neonatal mortality burden is still too high. Given the widespread recognition that perinatal outcomes are often misclassified and underreported, 23–25 and that there are still major gaps in stillbirth data in the UNHCR HIS, it is likely that neonatal mortality is underreported, and the true burden in refugee camp populations may be even higher that the upper bounds of figures presented here.
Even in countries hosting refugees in protracted camp situations, there has been limited change in the neonatal mortality burden over the last decade, although one would expect a decline over time. This may be due in part to limited attention to newborn health in humanitarian settings prior to the development of the Newborn Health in Humanitarian Settings Field Guide and UNHCR’s Operational Guidelines on Improving Newborn Health in Refugee Operations in 2014-2015.9,26 The across and within country variation in NMR, and lack of change over time, suggests that concerted efforts are needed to further understand and address neonatal mortality. UNHCR has already begun this work in some settings, including the establishment of a neonatal mortality audit system in Jordan, and projects focused on improving access to quality health services for women and newborns in Cameroon, Chad, Jordan, Kenya, Niger and South Sudan.27–30 Continued investment is needed in these and other refugee settings, including efforts to strengthen data availability and use as part of quality improvement efforts, as outlined in the recently launched Roadmap to Accelerate Progress for Every Newborn in Humanitarian Settings 2020-2024.31
Our analysis has several limitations. First, incomplete and/or inaccurate reporting of routine data and outdated population size estimates can lead to implausibly high or low mortality estimates. In some cases, high average NMRs over the study period may be explained by low numbers of births reported in certain years. For example, low camp refugee population and live births numbers in Burkina Faso might explain NMRs reported in 2015–2017. Second, although UNHCR guidelines suggest that all deaths occurring within a camp should be reported in the HIS, stillbirths and neonatal deaths taking place outside of health facilities may be underreported. Third, the UNHCR HIS does not differentiate between early or late neonatal deaths, or document cause of death, all of which are critical for identifying and addressing gaps in accessibility and quality of care. Finally, population-based survey data for sub-national divisions where refugee camps are located also have limitations and cannot be directly compared to routine health information system data.32 Surveys typically use a recall period of up to 2 years prior to survey administration; because availability of refugee camp HIS data varied from year to year, we compared NMRs reported in population-based surveys to NMRs reported in the UNHCR HIS in the year the survey was published, not the years where reported deaths occurred.
Data issues identified through this analysis point to opportunities for strengthening the UNHCR HIS to better facilitate monitoring and promote accountability for efforts to improve newborn health. These include increasing attention to the documentation of stillbirths in health facility registers and HIS reports; strengthening community health worker engagement in identification and reporting of stillbirths and newborn deaths occurring outside of health facilities; establishing mechanisms to distinguish between missing data and reports of zero deaths in a given month; incorporating alerts to flag potentially implausible ratios among neonatal, infant and child mortality in HIS; and consistently reporting neonatal mortality rates alongside infant and child mortality in public health reports. World Health Organization guidance cautions against calculating case fatality rates for time periods when the number of deaths are too small for a stable calculation.33,34 It may be advisable for facilities with few deaths to calculate NMR on a quarterly or even annual basis for increased stability of the indicator. Future questions to be researched include at which levels of mortality is a stable rate produced, how frequently indicators should be calculated, and how to account for factors unique to refugee camp settings, including in/out migration, community and health facility consolidation during time periods selected for calculation of NMR, which exacerbate challenges in obtaining accurate denominators for mortality rate calculation. Learning from these efforts may also inform efforts to strengthen routine health information systems and maximize the use of routine data for monitoring and evaluation of health development efforts.35–37 Outside of humanitarian settings, there are efforts to develop practical methods for improving measurements of perinatal mortality (fresh stillbirth and early neonatal mortality) in health facilities,38 and strengthen maternal and perinatal death surveillance and response systems which may provide models for replication in refugee camps.39–41