Mekelle: 13 April 2024 (Tigray Herald)
Neonatal mortality during the war in Tigray: a cross-sectional community-based study
preventable causes. An urgent response is needed to prevent the high number of neonatal deaths associated with the depleted health resources and services resulting from the war, and to achieve the Sustainable Development Goal on neonatal mortality.
Funding
UNICEF and United Nations Fund for Population Activities.
Introduction
The neonatal period is the most important time in establishing a child’s lifelong health and survival.1 For this reason, the world is striving to reduce the neonatal mortality rate to the Sustainable Development Goal of fewer than 12 deaths per 1000 livebirths by 2030.2, 3 Despite the remarkable decline of neonatal mortality in the past two decades, millions of neonates are still dying before age 1 month.4 Neonatal mortality is particularly a public health challenge in low-income countries, ranging from 25·7 deaths per 1000 livebirths to 79 deaths per 1000 livebirths.5, 6, 7 The burden is highest in sub-Saharan African countries,4, 7, 8 despite the global efforts to reduce this rate.9
Neonatal mortality in the Ethiopian Demographic Health Survey (EDHS) report has shown a declining trend for the past 15 years, from 49 deaths per 1000 livebirths in 2000 to 29 deaths per 1000 livebirths in 2016.10, 11 Many national and subnational estimates of neonatal mortality rates are reported in different studies. A neonatal mortality rate of 20·7 deaths per 1000 livebirths was also reported from the review of the EDHS 2016 data,12 which is close to the national target set by the Health Sector Transformation Plan II, to reduce neonatal mortality to 21 deaths per 1000 livebirths by 2025.13 Other recent studies estimated a neonatal mortality rate of 7·13 deaths14 and 19·8 deaths per 1000 livebirths for the Tigray region.15 These reports were done based on subnational estimates of the EDHS 2019 data. A government report from the Tigray region estimated a neonatal mortality rate of ten deaths per 1000 livebirths.16 Estimates of neonatal mortality in the country, and in Tigray, have been reported in a stable setting, indicating the need for data collection in a war setting.
Research in context
Evidence before this study
Neonatal mortality is among the key indicators of a country’s socioeconomic status. Information on neonatal mortality is crucial to designing interventions to protect against the preventable causes of newborn death. Neonatal mortality was 10 deaths per 1000 livebirths, according to the most recent study conducted by the Tigray statistical agency before the war in Tigray, Ethiopia, in 2020. Birth asphyxia, low birthweight, prematurity, and sepsis were known to be the leading causes of death globally. However, information on the magnitude of neonatal mortality and causes of neonatal death following the war in Tigray have not been investigated. We searched PubMed, Google Scholar, Web of Science, Popline, Embase, and WHO’s WHOLIS database for grey literature between Oct 23, 2022, and March 4, 2023. ”Neonatal mortality’’, ”neonatal death’’, ”early neonatal death’’, ”late neonatal death’’, and ”verbal autopsy’’ were used as search terms. Our search was limited to materials published in English.
Added value of this study
This study provides the most up-to-date information on the magnitude and causes of neonatal mortality during the war. It is also one of the first and largest studies to determine the magnitude and causes of neonatal mortality using verbal autopsy in the region, involving 189 087 households. The study has also informed the pattern of change in causes of death. The study reported a high rate of neonatal mortality, and asphyxia and infection were the leading causes of death in the early (0–6 days) and late (7–27 days) neonatal periods, respectively, which reflects the importance of reinstating lifesaving interventions as soon as possible.
Implications of all the available evidence
The findings of this study indicate the need for an urgent response to address the preventable causes of neonatal death following the war. The high neonatal death rate reflects a lack of access to emergency health care and medicines due to the effects of the war. The study provides information for researchers, decision makers, and programme managers to prioritise interventions, mobilise resources, and design future plans to reduce neonatal deaths and control the preventable causes of neonatal deaths during this conflict.
The leading causes of neonatal death globally are prematurity, birth asphyxia, and infections, which account for nearly 80% of deaths.17, 18 According to a facility-based study,19 these leading causes of death are echoed in Tigray. Another facility-based study in Tigray reported congenital anomalies as a leading cause of neonatal mortality, in addition to these common causes globally.20 Moreover, delayed health-care seeking, poor knowledge of the danger signs of newborns, a lack of trained health-care providers, and lack of essential drugs and equipment have also been reported as important causes of neonatal mortality in Africa.21, 22, 23, 24 In response to the high toll of neonatal mortality, various interventions have been implemented globally, including kangaroo mother care, community-based newborn care, and integrated community case management of newborn and childhood illness.5, 25, 26, 27
Fragile health systems and a disproportionally high mortality rate occur following war and conflict.28 In such situations, most deaths occur at home, where causes of neonatal mortality are rarely reported, especially in resource-limited settings.29 In conditions where civil registration of vital statistics is weak and the health system has collapsed, verbal autopsy remains the best option to determine the causes of death.30
In Ethiopia’s Tigray region, health promotion and disease prevention interventions were widely used before the war. Health facilities were made accessible to remote areas and ambulances were sufficiently available.31 However, 6 months into the war, only 17% of health centres, 30% of hospitals, 12% of ambulances, and none of the health posts (community-based health facilities)remained functional. The health centres and hospitals that remained in operation faced severe shortages in medicines, equipment, and staff.32 Moreover, looting of medical supplies, displacement of the health taskforce, and blockade of humanitarian access were common during the Tigray war.33 Consequently, hundreds of thousands of people are left without access to health care due to the collapse of the health system, which primarily affects vulnerable populations, such as mothers and their neonates.
Because routine health reporting was disrupted due to the conflict in Tigray, the magnitude and causes of neonatal mortality during the war have not been well documented. Thus, we aimed to do a community-based survey on neonatal mortality, to fill the data gap that is critically needed to mobilise resources for the rehabilitation of the health system.
Methods
Study design and population
The study was conducted in six zones, in 31 districts (22 rural and nine urban), 121 randomly selected tabias (kebelles), and 189 087 households in Ethiopia’s Tigray region. The western zone of Tigray and a few districts bordering Eritrea were excluded and inaccessible due to security reasons. The duration of the study was from Nov 4, 2020, to May 30, 2022, and the neonatal mortality survey was conducted over the month of June, 2022.
The sample was selected using a multistage stratified cluster sampling technique. In each of the selected zones, districts were stratified by urban and rural settings. Of the available 84 districts, a random selection of 31 (37%) districts was made that represented 22 rural and nine urban districts. This rural–urban stratification was done based on the nearly 30% urban and 70% rural population distribution of the region projected for 2022.34 The term urban district was used to represent the sub-city or town administration. The number of tabias in each district was determined as follows: three tabias (kebelles) were randomly selected from districts with fewer than seven tabias; from districts with between seven and 16 tabias, five tabias were randomly selected; and seven tabias were randomly selected from districts having more than 16 tabias. In districts that had fewer tabias, all the tabias in the district were used. All households in the selected 121 tabias were included using the census method to identify all births and deaths during the specified period.
A community-based cross-sectional study was conducted to assess the magnitude and causes of neonatal mortality during the war in Tigray. The study population included all livebirths and neonatal deaths (an infant death within the first 27 days of life) in the households of selected tabias and districts of the Tigray region. Respondents to the verbal autopsy interview were household members aged 18 years and older who provided care during the final illness that led to death (in the absence of the mother).
Ethical approval was obtained from the Research and Community Service Ethical Review Board, College of Health Sciences, Mekelle University (Tigray region, Ethiopia). At the district level, permission to conduct the study was obtained from the District Health Office before data collection. Study participants provided verbal informed consent before the interview. The research team was oriented to maintain the confidentiality of the respondents. Instead of names, anonymous or code was used when recording information during the interview. Study participants were informed about the right to withdraw anytime during interview. Only the research team had access to the collected data, and the filled questionnaire was kept secured and locked.
Procedures
The data collection process had two phases in all the selected districts. In the first phase, a census of all livebirths and all neonatal deaths that occurred in the household during the same period (from Nov 4, 2020, to May 30, 2022) was recorded. This was followed by a verbal autopsy interview for households with deceased neonates in the specified period using the translated WHO 2022 verbal autopsy tool. Global Positioning System coordinates were also taken from each selected household. Respondents to the verbal autopsy interview were adult relatives, sisters, or mothers who were caregivers at the time of a final illness that led to death.
Data on the sex (male or female) of the neonates were collected via self-report of the respondents.
A team of 121 data collectors and 31 supervisors were trained for 7 consecutive days. The contents of the training included the objective of the study, the 2022 WHO verbal autopsy guide, informed consent, selection of verbal autopsy respondents, how to sensitively approach verbal autopsy participants, ethical procedures, and general information regarding maternal death, including symptoms, medical history, and circumstances leading to death. Supervisors were responsible for checking the activity, consistency, and completeness of the questionnaire on the devices of the data collectors every day and providing overall support during the data collection process. Because the study involved a large population, additional follow-up supervision of the data collectors was done by the team of investigators (MME, HEA, HTe, AYL, BA, and KM) to ensure the quality of data collection in the sampled tabias, and timely feedback was given to the supervisors and data collectors. During data collection, migration uploading of data was not possible due to difficulties with internet access. Supervisors checked the quality of the data using the Open Data Kit in the field site, and the data were migrated to the server when the data collection period was over.
To aid interview participants, the verbal autopsy tool was translated into the local language (Tigrigna), and then translated back into English by language and health-care experts for consistency. The tool was validated in districts that were not selected for the study, following which modifications were made to the final version of the tool.
Data analysis
Data were collected using the Open Data Kit (version 1.27.3) and exported into STATA (version 15) for analysis. Descriptive data analysis was done using frequencies and percentages for the categorical data and median and IQR if quantitative variables were not normally distributed. The verbal autopsy data was processed using the Inter-VA-4 model (version 5.1) to assign the cause of neonatal death. The neonatal mortality rate was calculated per 1000 livebirths and stillbirth rate per 1000 births. An Inter-VA-4 computerised model was used to determine the likely cause of death.
Role of the funding source
The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
Results
29 761 livebirths were recorded during the screening of 189 087 households who agreed to participate in the census in the first stage of the study (figure 1). 317 neonates were stillborn, and 841 neonatal deaths were identified in 1158 households over an 18-month period from Nov 4, 2020, to May 30, 2022, in the 31 study districts. The neonatal mortality rate was 28·2 deaths per 1000 livebirths, and the stillbirth rate was 10·5 stillbirths per 1000 births.
A summary of the sample characteristics is presented in table 1. The mean age of the deceased neonates was 2·83 days (SD 5·4). 546 (65%) of the 841 neonatal deaths were male babies and 295 (35%) were female babies. Home was the place of death for 476 (57%) of the deceased neonates, with health facilities being the next most prevalent place. 673 (80%) of the deceased neonates resided in rural districts. 663 (79%) of 841 neonatal deaths occurred in the early (0–6 days) neonatal period.
The place of neonatal death stratified by age at death and by urban or rural residence is shown in table 2. Home was the most common place of death for early (54% [360 of 663]) and late (65% [116 of 178]) neonatal deaths. The health facility was the most common place of death after the home for both early (37% [245 of 663]) and late (29% [51 of 178]) neonatal deaths. Most (77% [130 of 168]) neonatal deaths in urban areas occurred in health facilities, whereas most (66% [446 of 673]) neonatal deaths in rural areas occurred at home.
Of the early neonatal deaths, most occurred within the first 24 h (appendix 2 p 1). As shown in appendix 2 (p 1) the death rate decreases with increasing age of the neonate remarkably.
The districts with the highest mortality rates reported repeated attacks (appendix 2 p 2). The rural districts experienced regular armed conflict that reduced residents’ ability to access health services. The health facilities in these districts were often not functioning due to damage and displacement of health workers.
The causes of death by neonatal period (ie, early vs late [7–27 days]) are shown in figure 2. The leading causes of death in 834 neonates (0–27 days) whose caregivers agreed to use of the verbal autopsy tool were birth asphyxia (35% [n=291]), prematurity (30% [n=247]), infection (12% [n=834]), and congenital malformation (4% [n=37]). The four leading causes of death accounted for more than 80% of neonatal deaths. In the late neonatal period, the leading causes of death were infection (28% [50 of 178]) and asphyxia (26% [46 of 178]); and in the early neonatal period, asphyxia (37% [246 of 663]) and prematurity (32% [214 of 63]) were the leading causes of death.
The causes of neonatal death by place are shown in figure 3. Home was the most common place of death for neonates who died from all causes except congenital malformations. Infection was a substantially more prevalent cause of death among those who died at home compared with those who died in health facilities. Most neonatal deaths at health facilities were due to congenital malformations, with birth asphyxia and prematurity as the next most common causes.
The causes of neonatal death by place are shown in figure 3. Home was the most common place of death for neonates who died from all causes except congenital malformations. Infection was a substantially more prevalent cause of death among those who died at home compared with those who died in health facilities. Most neonatal deaths at health facilities were due to congenital malformations, with birth asphyxia and prematurity as the next most common causes.
Discussion
This study provides information on the prevalence and causes of neonatal mortality during the war in Tigray. A high number of neonates in the Tigray region are dying within their first month of life, with most of these deaths occurring in the early neonatal period versus in the late neonatal period. The study identified the four leading causes of neonatal death, namely birth asphyxia, prematurity, infection, and congenital anomalies.
In the current study, neonatal mortality was found to be 28·2 deaths per 1000 livebirths, which is two to three times higher than that in previous studies conducted at the regional and national levels.12, 14, 16, 35 The neonatal death estimates range from 20·7 deaths per 1000 livebirths in 2016 to 7·13 neonatal deaths per 1000 livebirths in 2019. These studies were conducted in non-conflict settings when the basic maternal and newborn health services were more accessible, unlike in the context of the current study. The disruption of life-saving health services is likely to have increased the neonatal death rate relative to previous estimates because health facilities and the health workforce have been targets of the war.31, 36 However, the proportion of neonatal mortality could be underestimated in verbal autopsy studies: it is estimated that the misclassification rate of neonatal deaths as stillbirth ranges from 25% to 58%.37, 38, 39, 40 This type of misclassification is common in verbal autopsy studies (because most neonatal deaths occur at home) and might be biased due to misclassification of early neonatal death as stillbirth. Facility deaths are likely to distinguish between stillbirth and neonatal death.37
Neonatal mortality was much higher in rural areas than in urban areas in the current study, which is in line with results from previous studies.41, 42 Access to health services was poorer in the rural areas due to the damage of health infrastructure and displacement of health workers. Most of the maternal and newborn health services in the rural areas were disrupted, which is likely to be responsible for the high number of neonatal deaths.31 More than 50% of the deaths recorded in this study occurred at home, which is consistent with another study in which at least half of the neonatal deaths occurred after home births.43 There was also a high number of neonatal deaths that occurred in health facilities, which might be related to the lack of medical supplies and equipment.31
Neonatal mortality rate is among the common-use indicators of health service.1 The first week after birth was the most crucial time for a child’s survival and the period where neonatal deaths were most prevalent. 79% of neonatal deaths in our study occurred in the first week of life, which is higher than the global estimate of 75%44 and the WHO report estimate.18 The higher proportion of neonatal deaths in the current study are likely to be related to the health, social, and economic crises that occurred during the war in Tigray.
Birth asphyxia was the leading cause of early neonatal death in the current study, unlike previous reports of causes (including one from Yemen), which found that the leading cause of neonatal mortality was prematurity.7, 45 Birth asphyxia most often occurs due to delayed delivery care and is more likely to occur when there is a lack of access to health services during labour, delivery, and the early postnatal period. The availability of immediate postpartum care and essential newborn packages determines the survival of newborns with asphyxia and prematurity.4, 29 The scarcity of resources in war-stricken areas is often accompanied by overcrowding, poor staffing levels, and difficulty in providing even basic supportive treatment, which often results in high levels of neonatal mortality.46
We found that infection was among the most common causes of death in the late neonatal period. This finding is consistent with a study conducted in Rwanda, where sepsis and other infections played a major role in neonatal deaths after the first week of life.38, 47 Because most deliveries were at home, the less hygienic delivery process and scarcity of essential lifesaving medicines might predispose the newborns to infection48 and result in preventable death of the neonate. The survival of neonates requires access to health services including the availability of medical supplies.
To the best of our knowledge, this study is one of the first and largest studies to determine the magnitude and causes of neonatal mortality in Tigray. An Inter-VA-4 computerised model was used to determine the likely cause of death, which allows a standardised interpretation of the cause of death and is useful for processing large data sets.
Although the WHO 2022 verbal autopsy tool was used, the cause of several neonatal deaths was undetermined. Because interviews were held over an 18-month period, a potential recall bias can occur, particularly regarding the signs and symptoms of illness that led to the neonates’ time of death. The number of neonates who died in the first few hours after birth might be underreported for social or cultural reasons. As suggested by validation surveys, neonatal deaths are often underestimated in low-income settings due to misclassification errors as stillbirths.39 The high neonatal mortality rate in the current study might not be comparable with previous reports due to inconsistencies in the method of analysis, setting, and study period. This study evaluated the deaths of neonates; however, the future of the surviving neonates who could face starvation and lack of medication following the war requires further investigation.
Neonatal mortality is high in Tigray, which is likely to be partly attributable to the disrupted health service related to the war. Birth asphyxia, prematurity, and infection were the leading causes of neonatal death, and many of these cases are probably associated with poor access to quality health care. While the war remains ongoing, and in its aftermath, focused interventions in the intrapartum period will be required to improve the survival of neonates by improving access to quality maternal and newborn care.
Contributors
MT and AYL conceptualised the study and developed the protocol together with GF, HTe, and BB. MYH, TH, and GG supervised the data collection process during the field site. HEA, HTs, GF, and MME were responsible for data analysis including the data cleaning and validation. MT, BA, TsG, AM, and HG drafted the manuscript. TeG, AG, KM, and MG revised the manuscript and were responsible for journal selection and submission. All authors had access to all the data in the study and had final responsibility for the decision to submit for publication. MT, AYL, HTe, and GF accessed and verified the data and were also responsible for the decision to submit for publication. All authors reviewed and approved the final version of the manuscript.
Data sharing
The data are not publicly available. Data can be accessed from the principal investigator through formal request for research purposes. Please contact mache.tsadik@mu.edu.et to seek approval for data access.
Declaration of interests
We declare no competing interests.
Acknowledgments
This study was funded by UNICEF and United Nations Fund for Population Activities. We thank the study teams at Mekelle University, Tigray Health Bureau, and Tigray Health Research Institute (Mekelle, Tigray, Ethiopia). We also thank the supervisors, data collectors, and study participants.
Editorial note: The Lancet Group takes a neutral position with respect to territorial claims in published text, tables, and institutional affiliations.
and symptoms of illness that led to the neonates’ time of death. The number of neonates who died in the first few hours after birth might be underreported for social or cultural reasons. As suggested by validation surveys, neonatal deaths are often underestimated in low-income settings due to misclassification errors as stillbirths.39 The high neonatal mortality rate in the current study might not be comparable with previous reports due to inconsistencies in the method of analysis, setting, and study period. This study evaluated the deaths of neonates; however, the future of the surviving neonates who could face starvation and lack of medication following the war requires further investigation.
Neonatal mortality is high in Tigray, which is likely to be partly attributable to the disrupted health service related to the war. Birth asphyxia, prematurity, and infection were the leading causes of neonatal death, and many of these cases are probably associated with poor access to quality health care. While the war remains ongoing, and in its aftermath, focused interventions in the intrapartum period will be required to improve the survival of neonates by improving access to quality maternal and newborn care.
Contributors
MT and AYL conceptualised the study and developed the protocol together with GF, HTe, and BB. MYH, TH, and GG supervised the data collection process during the field site. HEA, HTs, GF, and MME were responsible for data analysis including the data cleaning and validation. MT, BA, TsG, AM, and HG drafted the manuscript. TeG, AG, KM, and MG revised the manuscript and were responsible for journal selection and submission. All authors had access to all the data in the study and had final responsibility for the decision to submit for publication. MT, AYL, HTe, and GF accessed and verified the data and were also responsible for the decision to submit for publication. All authors reviewed and approved the final version of the manuscript.
Data sharing
The data are not publicly available. Data can be accessed from the principal investigator through formal request for research purposes. Please contact mache.tsadik@mu.edu.et to seek approval for data access.
Declaration of interests
We declare no competing interests.
Acknowledgments
This study was funded by UNICEF and United Nations Fund for Population Activities. We thank the study teams at Mekelle University, Tigray Health Bureau, and Tigray Health Research Institute (Mekelle, Tigray, Ethiopia). We also thank the supervisors, data collectors, and study participants.
Editorial note: The Lancet Group takes a neutral position with respect to territorial claims in published text, tables, and institutional affiliations.
