Time to neonatal mortality and its predictors among preterm neonates admitted to the neonatal intensive care unit in northern Ethiopia, 2023/2024: a retrospective cohort study

Abstract

Background

A preterm neonate is defined by the World Health Organization as a child delivered before 37 weeks of gestation. In low- and middle-income countries, including Ethiopia, preterm-related complications are serious health problems due to increases in the mortality and morbidity of newborns and children under 5 years of age. The aim of this study was to assess the time to neonatal mortality and its predictors among preterm neonates admitted to the neonatal intensive care unit in northern Ethiopia, 2023/2024.

Methods

An institution-based retrospective cohort study was conducted among 495 randomly selected preterm neonates in six out of the fourteen general hospitals of Tigray, Ethiopia from October 2023 to June 2024. Epi Data version 4.6 and STATA version 14 were used for data entry and analysis, respectively. Descriptive statistics were carried out to determine the distribution. Kaplan-Meier analysis, life table, and log rank were computed. Cox proportional hazards models were fitted to identify independent predictors of preterm mortality.

Results

The proportion of preterm neonatal mortality was 109 (22.7%). The overall median survival time was 21 (95% CI: 20, 28) days. Initiation of breast milk (AHR = 0.38 (95% CI: 0.24, 0.61)), respiratory distress syndrome (AHR = 1.9 (95% CI: 1.07,3.63)), perinatal asphyxia (AHR = 2.05 (95% CI: 1.05, 4.00)), receiving kangaroo mother care practice (AHR = 0.5 (95% CI: 0.34, 0.83)), and gestational age (AHR = 1.6 (95% CI 1.07, 2.59) were the predictors of time to death.

Conclusion

Respiratory distress syndrome, gestational age less than 32 weeks, and perinatal asphyxia at admission were found to be independent risk factors for preterm neonatal mortality. Breastfeeding and receiving kangaroo-mother care were independent preventive predictors of preterm neonatal mortality. It is better to give full emphasis and close follow-up to preterm neonates, especially during the early neonatal period.

Text box 1. Contributions to the literature
• Providing a comprehensive analysis of time to neonatal mortality and its predictors among preterm admitted to Neonatal Intensive Care Units (NICU) in northern Ethiopia.
• Contributing to the understanding of neonatal mortality patterns in low-resource settings.
• Provide a holistic understanding of the factors that influence preterm neonatal survival in the NICU.
• Informing the development of targeted interventions to improve the survival of preterm neonates in northern Ethiopia.
• Raise awareness among healthcare providers and the public about preventable risk factors for preterm birth and neonatal mortality.

Background

A preterm neonate is defined by the World Health Organization (WHO) as a child delivered before 37 weeks of gestation or within 259 days of the start of a woman’s last menstrual cycle [1]. Preterm neonatal mortality, which begins at birth and lasts for the first month (28 days) of life, is the death of a preterm newborn [2]. Preterm neonates are more likely to experience complications such as Respiratory Distress Syndrome (RDS), hypothermia, Prenatal Asphyxia (PNA), neonatal sepsis, neonatal jaundice, and feeding issues because their physiological compensatory responses to the extrauterine environment are immature and insufficient [3].

Preterm delivery is the second-leading cause of death in children under the age of five worldwide and the main cause of neonatal mortality [2]. Approximately 290,000 neonates per year in sub-Saharan African countries lose their lives as a result of issues related to preterm birth, which accounts for approximately half of neonatal mortality in Africa. In addition to India, Nigeria, and Pakistan, Ethiopia has the fourth-highest neonatal mortality rate in the world [4]. Infection (24%) and problems connected to intrapartum (28%) and premature labor (37%) account for almost 90% of neonatal mortality fatalities in the nation [5]. Neonatal mortality decreased from 39 to 29 between 2005 and 2016, according to the Ethiopian Demography Health Survey (EDHS), but it increased to 30 in 2019 [6].

In 2019, there were 2.4 million neonatal deaths worldwide. Among these deaths, one-third of all neonatal deaths occur within 24 h of life, and close to three-quarters occur within the first seven days of life [7]. Preterm neonates born in sub-Saharan Africa are 10 times more likely to die than preterm neonates born in any developed country. Furthermore, preterm neonates born in sub-Saharan Africa are 12 times more likely to die than preterm neonates born in Australia and New Zealand [8, 9]. At 28 deaths per 1,000 live births, Ethiopia has a high neonatal mortality rate, with prematurity accounting for one-third of these deaths [10].

The neonatal phase, which lasts from birth to 28 days of age, is the most susceptible phase in terms of survival and health [11]. Each year, an estimated 14.8 million babies are thought to be born prematurely [12]. Of these, approximately a million pass away as a result of preterm problems, and 912 000 suffer from mild to severe neurodevelopmental impairments [13].

According to the EDHS report from 2016, one out of every 35 children died during the neonatal period [14]. Therefore, in 2005, the Ethiopian Ministry of Health (EMOH) created the first elaborate National Child Survival Strategy (2005–2015), which was put into practice as part of the third and fourth Health Sector Development Program (HSDP) cycles. The nation has recently updated its long-term plan (2015–2020) for child survival with the goal of eradicating all preventable child deaths by 2030 [15]. Preterm complications lead to consequences, such as significant medical costs for care, hospital stays, problems such as hearing loss and neurodevelopment defects, and the nation also loses healthy manpower [16].

According to a study performed in Ethiopia, 11.4% of preterm neonates admitted to Neonatal Intensive Care Units (NICUs) perished in the first 24 h and 85.27% died in the first seven days; therefore, these results suggest that the first few days are the most crucial for the survival of preterm newborns [10].

Ethiopia has created a variety of policies and programs, including NICU expansion, integrated neonatal and pediatric illness management, and a quality improvement program, to reduce newborn mortality at the institutional and community levels by preventing serious neonatal complications. Despite these measures, preterm neonatal death remains a problem [17]. Preterm mortality in Ethiopia is estimated to range from 14.6 to 35%, according to various studies [18, 19]. Particularly in Tigray, two retrospective cohort studies reported mortality rates of 14.6% and 32.1%, in two comprehensive specialized hospitals in northern Ethiopia [18, 20]. Because of the war between Tigray and Ethiopia, the severity of the problem is great, especially in Tigray, where the health system has collapsed [21].

The findings of this study aim to provide a holistic understanding of the factors that influence preterm neonatal survival in the NICU. This study also will be used by policymakers, and other stakeholders must be aware of the precise moment when there is a risk of preterm death to respond appropriately. To avoid and manage preterm issues, it is also important to identify predictors of death in relation to time, which will help develop the necessary training programs for healthcare professionals. Furthermore, this study was used to create awareness among health professionals and the community about how to reduce the avoidable risk factors that contribute to preterm mortality. It also serves as a source for future research and meta-analysis.

In Ethiopia, few previous studies have been performed to determine the predictors of preterm neonatal mortality. Nevertheless, the problem is still present. In addition, there are limited data, particularly in the Tigray region before the war. Owing to the war and harm caused to the healthcare sector, supporting data are lacking. Therefore, the aim of this study was to assess the estimation of time to neonatal mortality and its predictors among preterm neonates admitted to the NICU in general public hospitals of Tigray.

Methods

Study setting and design

An institution-based retrospective cohort study design was conducted in the Tigray region, which is one of the 11 federal administrative regions in Ethiopia. It is the homeland of the Tigrayan, Irob, and Kunama people, with the capital city of Mekelle. The study was conducted from October 2023 to June 2024.

Inclusion and exclusion criteria

All live preterm neonates admitted to the NICU in selected general hospitals in Tigray during the data collection period were included. All preterm neonates with charts with incomplete records or missing important variables (outcome of status, date at which outcome was determined, and duration of stay) were excluded from this study.

Sampling

The sample size was determined via STATA (Cox model) version 14 with the following assumptions: two-sided significance level (α) of 5% =0.05, power 80% = 1.28, Za/2 = Z value at 95% confidence interval = 1.96, mortality rate = 14.6%. To estimate the sample size, information on significant predictors of time to among preterm neonates’ neonatal mortality was used. Among these predictors we used the fetal compound presentation, which refers to the fetal presentation in which an extremity is present alongside the part of the fetus closest to the birth canal, given it had the strongest association with the outcome Adjusted Hazard Ratio (AHR) of 2.29 and provided a maximum sample size [18]. The design effect was multiplied by 1.5, and nonresponse rate (for compassion of incomplete cards) of 5% was added. The final sample size for this study included 495 preterm neonates admitted to the NICU and having a medical chart.

There are 14 general public hospitals in the Tigray region, six of which were selected by using a simple random sampling technique to conduct the study. These hospitals are Suhul General Hospital, St. Mariam General Hospital, Adigrat General Hospital, Mekelle General Hospital, Wkuro General Hospital, and Abyi Adi General Hospital. A simple random sampling approach was used to select the charts of preterm neonates from the logbooks or to register them via a computer-generated method in the selected general hospitals. The necessary number of samples was proportionally distributed among all of the general hospitals that were chosen.

Data collection tools and techniques

An appropriate data collection tool was adapted from the literature and WHO guidelines and prepared in English. The data abstraction checklist comprises baseline sociodemographic characteristics of both the neonate and the mother, maternal medical disorders, obstetric related factors, common medical disorders in the neonate, date of admission, date of discharge, and outcome of preterm birth. The preterm neonatal chart number was taken from the medical record book of the NICU. Before the data were collected, the chart was reviewed (baseline and follow-up records). The records of preterm neonatal charts were selected according to the eligibility criteria.

Data were collected by data collectors via a data abstraction checklist. It was retrieved by 4 neonatal nurses and 2 BSc. nurse supervisors. The starting point for this study was the first date of admission to the NICU, and the end was the date of death, censored (date of discharge, referred, or alive at the end of the study) until the last 28 days. The time to death and mortality of preterm neonates were extracted from the medical records of preterm neonates who were admitted to the NICUs. The data collection was conducted from November 24 to December 24, 2023.

Data quality control

To ensure data quality, a pretest was performed on 5% of the data abstraction checklist from the total sample size one week prior to the actual study to check the consistency and quality of the checklist. Modifications were accordingly made after the pretest. Three full days of training were given to the data collectors and supervisors regarding the study, the checklist, and the data collection procedure by the main investigator. The collected data were checked every day by the supervisor and principal investigator for completeness. The problems faced were discussed over the night with the data collectors and the supervisor, and they were kept in the form of a file in a secure place where no one could access it except the investigator. Confidentiality was ensured by not recording names or any personal identities. Data cleanup was performed by running the frequencies of each variable to check for accuracy, missing data, and consistency.

Data processing and analysis procedures

Before analysis, the data were coded, edited, and cleaned via the Epi Data statistical software package version 4.4.2.1, and any errors identified at that time were corrected after the review of the original data using the coded numbers. The data were subsequently exported to STATA version 14 for analysis. Descriptive statistics were carried out to determine the frequency and percentage of the dependent and independent variables. The mean ± SD are presented for normally distributed continuous covariates, whereas the median and Inter-Quartile Range (IQR) are presented for skewed covariates. The outcome was dichotomized as censored or dead. The incidence rate was also calculated by the person’s time of observation.

The time to mortality of preterm neonates was calculated as the time between the date of admission and the date of death, censored, or the end of the admission period. Kaplan‒Meier analysis was also used to estimate the median survival time, and the cumulative probability of survival and to compare the survival differences between the different covariates. The log rank test is also used to compare significant survival differences between categories of different explanatory variables. A life table was used to estimate the cumulative probability of survival at the different time intervals. The proportional hazard assumption was verified for both the bivariable and multivariable variables. The Cox regression model was applied to describe the associations between the dependent and independent predictors of preterm mortality.

To control for possible confounding covariates simultaneously, the covariates that had a p value < 0.25 in the bivariable analysis were transferred to the multivariable analysis, those variables with a p value < 0.05 at the 95% confidence level were considered independent predictors of preterm mortality, and the variance inflation factor was used to check for multicollinearity among the independent variables. The crude hazard ratio and adjusted hazard ratio were used to test the strength of association-dependent and independent variables. The Cox proportional hazard regression assumption was tested by using the Schoenfeld residual test. The overall goodness of fit of the model is would be checked graphically via the Cox–Snell residual graph. Finally, the results are presented in tables, figures, and text accordingly.

Results

Out of the 495 selected preterm neonates, the medical charts were complete resulting in a 97% response rate. They had different lengths of hospital stay, with a minimum of 1 day and a maximum of 28 days, with a median follow-up period of 5.5 days and a mean (SD) length of hospital stay of 6.6 days, which resulted in 3,178 person-days of observation overall.

Preterm neonate mortality

In this study, 109 (22.7%, 95% CI: 19.16, 26.68) preterm neonates died during the admission period. About 371 (71.3%) neonates were censored; 15 were referred to another facility (3.13%), 9 (1.88%) discharged against medical advice, and 347 (72.29%) were discharged (Fig. 1). The median length of follow-up was 5.5 days. The total person-day observations were 3,178 days. The overall incidence rate of mortality for preterm neonates was 34.2% (95% CI: 28.4, 41.3) per 1,000-person days. Furthermore, the overall median survival time of preterm neonates was 21 (95% CI: 20, 28) days.

Fig. 1

Overall survival status of preterm neonates admitted at the NICU, northern Ethiopia

Sociodemographic predictors

In this study, approximately 480 recorded medical charts were obtained and reviewed at six hospitals. Preterm neonatal medical records were taken from each hospital on the basis of previous admissions: Mekelle general hospital, 73 (15.21%); Adigrat general hospital, 83 (17.29%); St. Mariam general hospital, 27 (5.63%); Suhul general hospital, 119 (24.79%); Abyi Adi general hospital, 107 (22.29%); and Wkuro general hospital, 71 (14.79%). Among these hospitals, approximately 109 (22.7%, 95% Cl: 19.16, 26.68) preterm neonates died. Approximately 290 (60.4%) preterm neonates were males; among these, 65 (22.4%) died. About 280 (58.3%) mothers were from rural residences; of those, 73 (26.1%) died. Only 70 (14.6%) preterm neonates were born at home, and 12 (17.1%) of those died. The median maternal age was 26 year with an IQR of (24–29) years (Table 1).

Table 1 Distribution of sociodemographic characteristics of mothers and preterm neonates admitted to the neonatal intensive care unit of general hospitals in Tigray, Northern Ethiopia, 2023/2024

Maternal medical diagnosis and obstetric related predictors

Among the 480 preterm neonates, 69.2% were born to mothers who had a parity of less than 2, and 399 (83.1%) were born to mothers who had antenatal care (ANC) follow-up. Of these, 84 (21.1%) died. A total of 123 (30.7%) babies were born to mothers who had three or more ANC visits. A total of 51 (10.6%) were born to mothers who had obstetric complications; among these, 22 (4.6%) were born to those who suffered from eclampsia; out of them, 11 (50%) died. Of the 480 preterm neonates, 433 (90.2%) had cephalic presentation, and almost all of them were delivered by spontaneous vaginal delivery 393 (81.9%). There were six preterm neonates delivered from mothers with HIV/AIDS, three of whom died (Table 2).

Table 2 Obstetric and medical characteristics of mothers admitted to the neonatal intensive care unit of general hospitals in Tigray, Northern Ethiopia, 2023/2024

Neonatal and preterm birth–related predictors

In this study, 313 (65.2%) preterm neonates had birth weights between 1,500 and 2,500 g; 43 (13.7%) died. A total of 128 (26.7%) preterm neonates did not receive any breastmilk after delivery; 59 (46.1%) of those died. Moreover, 127 (26.5%) neonates did not receive Kangaroo Mother Care (KMC). Among these preterm neonates, 57 (44.9%) died. There were 104 (21.7%) neonates whose APGAR score with in the 1st minute was less than seven, of whom 39 (37.5%) died. A total of 312 (65.8%) moderate-to-late preterm neonates were born; among these, 50 (14.7%) died. The median gestational age at birth was 34, with an IQR of (32–35) weeks. The overall incidence rate of mortality for preterm neonates was 34.2 (95% CI: 28.4, 41.3) per 1,000-person days. Only 23 (4.8%) preterm neonates did not experience medical or surgical complications. Among those who had such complications, 146 (31.9%), 139 (29%), 65 (12.7%), 47 (9.8%), 34 (7.1%), and 26 (5.4%) had neonatal sepsis, RDS, PNA, hypothermia, hypoglycemia, and neonatal jaundice, respectively (Table 3).

Table 3 Distribution of neonatal and preterm birth related characteristics for preterm neonates admitted to the neonatal intensive care unit in general hospital in Tigray, Northern Ethiopia, 2023/2024

Overall survival function and life table

The overall Kaplan‒Meier survival estimate revealed that the probabilities of survival of preterm neonates at the end of the first day, 7th day of admission, 14th day of admission, 21st day of admission, and 28th day of admission were 97.5%, 74.3%, 62.7%, 47.6%, and 23.8%, respectively. This is the inverse of a hazard function. The survival time was a decreasing step function. Kaplan‒Meier survival function estimation graphs were prepared for categorical covariates to observe survival differences (Figs. 2 and 3).

Fig. 2

Overall survival time function for preterm neonates admitted in NICU, northern Ethiopia

Fig. 3

Overall hazard function for preterm neonates admitted in NICU, northern Ethiopia

The life table helps to show the survival probabilities of preterm neonates over an interval of time. The survival probabilities of preterm neonates at 0–7 days, 7–14 days, 14–21 days, 21–28 days, and 28 days were 79%, 64%, 55%, 40%, and 21%, respectively. As time increases, the death rate of preterm neonates increases, but their survival decreases (Table 4).

Table 4 Life table with seven-time intervals to death of neonates admitted to the neonatal intensive care unit in general hospitals in Tigray, Ethiopia, 2023/2024

Predictors of time to mortality in preterm neonates

Cox proportional hazard regression, both bivariable and multivariable, was fitted.

to identify predictors for the time to mortality of preterm neonates. Variables that had a p-value < 0.25 in the bivariable analysis were entered into the multivariable analysis. Findings from the bivariable analysis revealed that residence, neonatal sepsis, ANC visit, RDS, birth weight, perinatal asphyxia, kangaroo mother care, gestational age, APGAR score at 1 and 5-minutes, obstetric complications, and initial breastfeeding were associated with the time to mortality of preterm neonates. Multicollinearity was checked by the Variance Inflation Factor (VIF) with an overall score of 1.21 and multivariable to identify the variables that were independently significant. In the multivariable analysis, five variables, namely, RDS, perinatal asphyxia, kangaroo mother care, initiating breastfeeding, and preterm neonates with < 32 weeks of GA were identified as independent predictors of the time to mortality of preterm neonates (Table 5).

Table 5 Bivariate and multivariate Cox proportional hazard regression outputs for predictors of the time to mortality of preterm neonates admitted in neonatal intensive care unit general hospitals in Tigray, Northern Ethiopia, 2023/2024

Survival function and comparison of survival experience by PNA

Preterm neonates who had experienced PNA had a lower survival experience than their counterparts did (using the log rank test, X² = 9.36, P-value = 0.0022) (Fig. 4). The median survival time for preterm neonates who had experienced PNA was 12 days.

Fig. 4

Kaplan-Meier survival time of preterm neonate with prenatal asphyxia at NICU, norther Ethiopia

Survival function and comparison of survival experience by KMC practice

Preterm neonates who received KMC had greater survival experience than did neonates who did not receive KMC (using the log rank test, X² = 14.18, P-value = 0.0002) (Fig. 5). The median survival time for preterm neonates who had experienced KMC practice was 28 days.

Fig. 5

Kaplan-Meier survival time of preterm neonate having KMC practice at NICU, northern Ethiopia

Survival function and comparison of survival experience by initiation of breastfeeding

Preterm neonates who had experienced the initiation of breastfeeding had greater survival experience than did neonates who did not experience the initiation of breastfeeding (using the log rank test, X² = 67.0000, P-value = 0.0002) (Fig. 6). The median survival time for preterm neonates who had experienced the initiation of breastfeeding was 28 days.

Fig. 6

Kaplan-Meier survival time of preterm neonate with initiate breast feeding at NICU, norther Ethiopia

Assessing the proportional hazard assumption

The global test of the proportional hazard assumption based on the Schoenfeld residuals revealed that all of the independent variables had p values > 0.05, and the full model satisfied the Cox-proportional hazard assumption at an overall global test of p values = 0.38.

Goodness of fit test

The plot of Cox–Snell residuals is closest to 45⁰ straight lines through the origin for the Cox proportional hazard model compared with the parametric survival model. This suggested that the Cox-proportional hazard model provided the best fit for this dataset (Fig. 7).

Fig. 7

Model fit of the Cox-Snell residual test among preterm neonates admitted at NICU, northern Ethiopia

Preterm neonates with respiratory distress syndrome were 1.9 times (AHR: 1.9; 95% CI: 1.07–3.63) more likely to die than their counterparts. Preterm neonates with perinatal asphyxia were twice (AHR: 2.05; 95% CI: 1.05–4.00) more likely to die than their counterparts. A gestational age of less than 32 weeks increased mortality by 1.5 times greater than that at 32 weeks (AHR = 1.6; 95% CI: 1.07–2.59). Preterm neonates who had received kangaroo mother care practices were 47% (AHR: 0.52; 95% CI: 0.34–0.83) less likely to die than their counterparts were. Preterm neonates who initiated breastfeeding were 62% (AHR: 0.38; 95% CI: 0.24–0.61) less likely to die than their counterparts.

Discussion

According to the findings of this study, the proportion of preterm neonatal mortality was 22.7% (95% Cl: 19.16, 26.68). The overall incidence of mortality and the median survival time for preterm neonates were 34.2 (95% CI: 28.4, 41.3) per 1,000 person-day observations and twenty-one (95% CI: 20, 28) days, respectively.

The findings of this study are slightly higher than those of a study conducted in Ayder compressive specialized hospital, northern Ethiopia [18]. The reason might be the difference in study area coverage; sample size variation and the war in Tigray affect the health setting as well as health professionals [21]. On the other hand, this finding is slightly lower than those of studies conducted in Ethiopia at the University of Gondar Specialized Hospitals, Felegehiwot Specialized Hospital, Aksum and Ayder Comprehensive Specialized Hospital, Jimma University Specialized Hospital, and Mizan Tepi Hospital [10, 19, 20, 22, 23]. This variation could be due to the study’s time variation, area of coverage, or differences in population size used.

The finding from a study conducted in Uganda is lower than the finding in the current study [24]. These marked differences in the magnitude of preterm neonatal mortality might be due to differences in socioeconomic status and geographic location; changes in treatment modalities can be the reason for the observed differences. This finding is consistent with studies conducted at Tikur Anbesa Specialized Hospital, Jimma University Hospital, Debre Tabor Hospital, and Felegehiwet Hospital [22, 23, 25, 26].

In this study, the overall median survival time of death in preterm neonates admitted to the NICU was 21 days, which is higher than the study done at Mizan Tepi Hospital, Ethiopia [19]. The possible reason might be a disparity in study area and population size. In contrast, this study is in line with studies conducted at Tikr Anbesa Specialized Hospital and the University of Gonder Specialized Hospital [10, 25].

This study showed that RDS was an independently significant predictor of the time to mortality of preterm neonates. Preterm neonates who had experienced RDS were more likely to die compared to those who had not experienced RDS. This finding is supported by studies conducted in different parts of Ethiopia: University of Gondar specialized hospital, Debre Markos, Addis Ababa, Jimma University specialized hospital, Hawasa University, Felege hiwot hospital, Ayder comprehensive specialized hospital, Northern Ethiopia, and Aksum and Ayder comprehensive specialized hospital [10, 18, 20, 22, 23, 27,28,29].

Preterm neonates who received KMC had a lower hazard of preterm neonate mortality compared to preterm neonates who did not receive KMC. The result was consistent with the study conducted at Aksum and Ayder comprehensive specialized universities [20], Ayder compressive specialized hospital of Northern Ethiopia [18], and the University of Gondar specialized hospital [10]. This similarity might be due to the similarity of the study setting, study design, and KMC practice, which prevent the risk of hypothermia by decreasing body surface area to the external environment, preventing infection, improving gastrointestinal function and cardiorespiratory stability, and promoting or encouraging breastfeeding.

Preterm neonates who had PNA had a greater hazard of preterm neonatal mortality than those who had not PNA. This finding is consistent with studies conducted in Ethiopia, such as at the University of Gondar specialized hospital, Addis Ababa, Felegehiwot hospital, Ayder and Aksum comprehensive specialized hospital, and Jimma specialized hospital [10, 20, 22, 23, 27].

Initiation of breastfeeding was the other preventable variable that significantly predicted the time to mortality of preterm neonates. Preterm neonates who were initiated on breastmilk were less likely to die compared to those preterm neonates who were not initiated on breast feeding. This finding is supported by different studies conducted in the Ayder compressive specialized hospital of Northern Ethiopia, the University of Gondar Specialized Hospital, and the Jimma University Specialized Hospital [10, 18, 23]. From the scientific evidence, the initiation of breastfeeding might provide protection against various diseases since breastmilk contains antibacterial, immunologic, and other factors that can enhance bactericidal enzymes, complements, and macrophages, and help to prevent hypoglycemia and hypothermia.

Furthermore, gestational age was also another predictor of preterm neonatal mortality; the risk of death for preterm neonates who were < 32 weeks of GA at the time of admission was higher than that for those who were between 32 and 37 weeks of GA. This finding is supported by studies conducted at Jimma University specialized hospitals, the University of Gondar specialized hospital, and Mizan Tepi Hospital [10, 19, 23].

This study revealed that the mortality of preterm neonates on the first day was 2.5%, that on the second day was 5.6%, that on the seventh day was 25.6%, and that at 21 days was 52.3%. This result indicates that, from time to time, the mortality of preterm neonates has increased. This might be due to failure to detect complications early, failure to apply preventive strategies, safe quality care accordingly, and inability to perform timely intervention after the first day of life.

Strengths and limitations

This study has strengths, such as the fact that it was performed in different hospitals and that data were collected by primary-trained neonatology nurse professionals who are familiar with treatment guidelines and records of treatment at the NICU. This design helps to identify the temporal relationship of preterm neonatal mortality with predictor variables and considers time. This study has limitations, such as that some important predictors of preterm neonatal mortality, such as maternal educational status and maternal nutritional status, were not explored, nor were service-related predictors since the data were collected from secondary sources. While all preterm neonates were considered for inclusion, this study is limited to those admitted to the NICU specifically for preterm-related conditions.

Conclusion

In this study, respiratory distress syndrome, gestational age less than 32 weeks, and perinatal asphyxia at admission were found to be independent risk factors for preterm neonatal mortality. Breastfeeding and receiving kangaroo-mother care were independent preventive predictors of preterm neonatal mortality. Therefore, providing care with full emphasis and close follow-up for preterm neonates, especially during the early neonatal period, and strengthening obstetrics care to reduce prematurity, the use of antenatal steroids for women with imminent preterm delivery, neonate administration of dexamethasone to prevent immaturity of the lung, and early diagnosis and management of neonatal complications such as RDS, PNA, and others (Table 6).

Table 6 Identified preventable factors for neonatal mortality among preterm neonates admitted to the neonatal intensive care unit of general hospitals in Tigray, Northern Ethiopia, 2023/2024