Pattern, Causes, and Treatment Outcome of Neonates Admitted at Hiwot Fana Specialized University Hospital, Eastern Ethiopia.

Background: Neonatal period is the most the vulnerable time for survival of newborns. In Ethiopia, neonatal death remains increasing and requires extraordinary efforts and a novel intervention to reduced neonatal death as those deaths continue to occur in even hospital settings where there are the best neonatal care services. Despite many efforts done to improve the outcome of neonates admitted to hospitals; neonates continue dying in those settings. The causes of neonatal death in Ethiopia varies from place to place. So, identifying the causes of death in the study area is very important for prevention and treatment. The study aimed to assess the causes and factors associated with treatment outcomes. Methods: Institution-based cross-sectional study design was conducted among 707 randomly selected neonates from March 1 to 31, 2020. Data were extracted from medical records using a checklist adapted from WHO and neonatal registration book. The data were inserted into Epi-data version 3.1 and then exported into SPSS window version 22 for analysis. Bivariate and multivariate analyses were done to identify the association between independent variables and the outcome variable. Results: From 698 admitted neonates during 2 years period, 594 of them were improved and 104 of them were died. The proportion of neonatal death was 14.9% (95% CI:12.3,17.9). Neonatal sepsis, low birth weight and prematurity were the leading cause of neonatal death. Residency [ AOR=2.30, 95%CI: (1.3, 4.12)], low birth weight [AOR=2.52, 95%CI: (1.24,5.13))], respiratory distress syndrome [AOR =2.86, 95%CI: (1.11.7.35)], neonatal sepsis [AOR= 2.48, 95%CI: (1.40, 4.38)], and Neonates treated with phototherapy and oxygen [ AOR=0.22, 95%CI:(0.088,0.54)], [AOR= 0.47,95%CI:(0.22,0.99)] were factors associated with poor treatment outcome. Conclusion: The causes of death were mainly from causes like sepsis, low birth weight, and prematurity in the study setting. Residency, low birth weight, respiratory distress syndrome, having neonatal sepsis, treatment with phototherapy, and treatment with oxygen were independent factors. The concerned bodies should give attention to neonates admitted to intensive care units by strengthening the quality of care given at the unit and strengthening early detection and prevention of neonatal problems during post-partum periods.


Introduction
Neonatal period refers to the rst 28 days of life. The rst 7 days of life is termed as early neonatal period while the period between 7 days of life and 28 days is termed as late neonatal period [1]. Neonatal period is the most vulnerable time for the survival of newborns. Globally, 2.5 million neonates died in 2018.
Children continue to face widespread regional problems in their chances of survival. The highest rates of child mortality are still in Sub-Saharan Africa, where 1 child out of 9 children dies before age ve and this region is among the regions showing the least progress which accounts for 38 percent of global neonatal deaths [2].
Even though NICU has advanced technology and trained healthcare professionals to give special care for sick neonates and for those who need specialized nursing care they continue to die [3]. In Ethiopia, neonatal death was declined from 39 deaths per 1000 live birth in 2005 to 28 deaths per 1000 live birth in 2014. However, the recent report shows a rise in neonatal mortality from 28 death per 1000 live birth in 2014 to 30 neonatal death per 1000 live birth in 2019 [4]. It clear that the majority (82.4 %) of neonatal deaths occurred within the rst week of life [5] and requires exceptional efforts and a novel intervention to reduced neonatal death as this mortality continue to occur in even hospital settings where there are best neonatal care services.
Ethiopia has developed and adopted national, and international child health intervention strategies; including national newborn and child survival strategy (2015-2020) to reduce neonatal death through interventions including neonatal resuscitation, cord care, thermal regulation, Kangaroo Mother Care, antibiotics for neonatal sepsis, and early initiation of breastfeeding by 2020 [6,7]. Besides, Ethiopia is striving to achieve SDGs, aiming to end preventable deaths of newborns and children under 5 years of age by 2030 [8]. This warrants a study of factors that in uence the poor neonatal outcome in the hospital setting that would help guide clinical practices and policy recommendations to achieve the intended goals as a nation.
Even though there are variations concerning the causes of neonatal death, evidence reveals that infections, prematurity, and birth asphyxia are the leading causes of mortality [9]. It is important to measure the extent of neonatal death and identify in uencing factors that would guide the development of focused and evidence-based health interventions to reduce poor neonatal outcomes at the hospital. To our knowledge, this evidence is limited or inadequate therefore, this study aimed to assess the cause and factors associated with poor neonatal outcomes among neonates admitted in Neonatal Intensive Care Unit (NICU) at Hiwot Fana Specialized University Hospital (HFSUH), Harar, Eastern Ethiopia.

Study Setting and Period
This study was conducted at the neonatal intensive care unit of Hiwot Fana Specialized University Hospital (HFSUH), in Harar City. The city is located 526 km to the east of Addis Ababa, the capital of Ethiopia. HFSUH currently provides different services for approximately 5.8 million peoples in the catchment area. The NICU is one of the Intensive Care Unit (ICU) services rendered in the hospital. The hospital admits about 120 neonates per month on average. The unit is divided into a septic room, kangaroo mother care (KMC) room, and critical and subcritical rooms. The unit has 19 neonatal beds and 14 KMC beds, 5 incubators, 10 radiant warmers, and 4 phototherapy machines. Additionally, there are 8 infusers, 4 oxygen cylinders, pulse oximetry, glucometer, and neonatal resuscitation equipment. The unit is staffed with 6 pediatricians, pediatric residents, 4 neonatal nurses, and 7 clinical nurses. The study was conducted from March 1 to March 31, 2020.

Study Design and population
Institution-based cross-sectional study design was implemented. From a total of 2536 two years of admission (December 1, 2018, to December 31, 2020), 707 study units were selected randomly. Neonates with incomplete medical records, neonates, who were referred to other places and left against medical advice were excluded from the study.
Incomplete medical records. Neonates left against medical advice (self-discharge was excluded from the study (Fig. 1).

Data Collection Methods and Quality assurance
Data were extracted from medical records of newborns using a checklist adapted from the WHO document of review and audit of neonatal death [17] and registration book neonates. We recruited 10 nurses holding BSc degree and trained them to collect the data and the primary investigator supervised the process of data collection to ensure quality.
We extracted the age of mother, residence, ANC follow-up, parity, gravidity, number of pregnancies, neonatal illness, gestational age, birth weight, sex of neonate, APGAR score, treatment provided, treatment outcome, length of stay, place of delivery, and mode of delivery, Antepartum Hemorrhage (APH), postpartum hemorrhage (PPH).

Operational De nitions
Neonatal death is de ned as the number of neonates who died at the neonatal intensive care unit throughout the study period [10].
Hypothermia: Any low body temperature measurement (< 36.5°C) was diagnosed and recorded on charts during the admission of neonates [12].
Prematurity is described as live-born neonates delivered before 37 completed weeks that are already diagnosed by professionals in charge of the admission of the neonate to neonatal intensive care units [13].
Sepsis: Record of infection or sepsis diagnosed either clinically or with culture by professionals during admission of the neonate and recorded on the chart [12].
Birth asphyxia: is diagnosed whenever a neonate had an Apgar score < 6 in the fth minute and/or if he/she did not cry immediately after birth; had respiratory distress, oppiness, loss of consciousness, presence of convulsion, and loss of neonatal re exes [13].
Birth weight is classi ed using WHO weight classi cation, Low birth weight is any child with birth weight is less than 2,500 g [14]. All other assessments are based on physician judgment as written in the patient card.

Data processing and analysis
The data was coded, edited, cleaned, and entered into Epi data statistical software version 3.1 and then exported to SPSS version 20 for analysis. A descriptive statistical analysis was used to summarize data.
The information was presented using frequencies, tables, and gures.
Bivariate analysis and multivariate analysis were done to observe the association between independent variables and the outcome variable by using binary logistic regression. All variables with P ≤ 0.25 in the bivariate analysis were included in the nal model of multivariate analysis to control all possible confounders. The model goodness of t was tested by the Hosmer-Lemeshow statistic and Omnibus test.
The model was considered a good t since it is found to be insigni cant for Hosmer-Lemeshow statistic (p = 0.221) and signi cant for Omnibus tests (p = 0.000) The multi co-linearity test was carried out to observe the correlation between independent variables using VIF, and standard error, no variables were observed with VIF of > 10 and standard error > 2. The direction and strength of statistical association were measured by the odds ratio with 95% CI. The adjusted odds ratio along with 95% CI was estimated to identify factors for neonatal mortality by using multivariate analysis in the binary logistic regression. In this study P-value < 0.05 was considered to declare a result as a statistically signi cant association.

Result Maternal Socio-Demographic Characteristics
A total of 698 admitted neonates during study period were included for further analysis. The present study shows that more than half 365 (52.3%) of mothers of neonates came from the urban areas while 333 (47.7%) were from rural. Maternal age was documented for 612 (87.7%), of which 51 (6.9%) of mothers were below age 20. From 698 mothers their median age was 27 years with an interquartile range of 10 years.

Obstetrics Characteristics
More than three fourth (74.1%) of mothers had ANC follow-up and 23.6% of mothers had 4 and above ANC visits. There were 269 (38.5%) primiparous mothers and 325 (46.6%) were multiparous mothers. There were 40 (5.7%) mothers with a history of neonatal death. The majority of 639 (91.5%) mothers had singleton pregnancy. The majority 590 (84.5%) of the neonates were delivered at the hospital while 32 (4.6%) of neonates were delivered at home. Among 698 delivered neonates, 161 (23.1%) of neonates were delivered by Cesarean section. Six hundred fty-seven (94.1%) of neonates were born within 24 hours of the start of labor (Table 1).

Neonates Characteristics
Three hundred eighty-four (55%) neonates were males. More than two-thirds (69.1%) of neonates were admitted within 24 hours of life. The median age at admission was 9.5 hours with an interquartile range of 71 hours. Most 660 (80.4%) of neonates stayed at the hospital for less than 7 days. The median hospital stay was 72 hours with an interquartile range of 120 hours. APGAR score was documented for 674 (82.1%) of neonates. Out of those, a low APGAR score was recorded for 129 (19.1%) and 24 (3.6%) of neonates in the 1st and 5th minute respectively. Birth weight was documented for 613 (87.8%) neonates. Among those, one hundred eighty-six (30.3%) neonates were low birth weight, 39 (6.4%) neonates were very low birth weight and, 2 (0.3%) neonates were extremely very low birth weight. More than half (500, 71.6%) of neonates were term while 198 (29.4%) of neonates were preterm ( Table 2).  (Table 3).

Treatment given for Admitted Neonates
Antibiotics were given for more than half 485 (69.5%) of admitted neonates. The most frequently administered one being a combination of ampicillin and gentamycin 346 (71.3%). Other combinations were found to be ampicillin with ceftriaxone 54 (11.1%) and vancomycin with cefotaxime 31 (6.4%).

Pattern of neonatal mortality
More than three fourth of the deaths (76, 73%) occurred within the rst 24 hours of age and 11% of death occurs within 72 hours of age. The death pattern declined as the age of the child crossed 72 hours (Fig. 5).

Factors Associated with Treatment Outcome of Admitted Neonates
In the bivariate model, treatment outcome was signi cantly associated with residency, ANC follow-up, complication during pregnancy, complication during labor and delivery, place of delivery, prematurity, LBW, sepsis, perinatal asphyxia, pathologic jaundice, respiratory distress syndrome, APGAR score on the 1st minute, neonatal resuscitation, glucose, oxygen, phototherapy, and length of hospital stay.
To control the effect of confounders variables including residency, presence of ANC follow-up, complication during labor and delivery, place of delivery, APGAR score on 1st minute, prematurity, LBW, neonatal sepsis, RDS, neonatal resuscitation, phototherapy, oxygen, and glucose were selected for the nal model. Residency, LBW, neonatal sepsis, RDS, phototherapy, and oxygen were signi cantly associated with treatment outcome.
The multivariable analysis showed that the odds of neonatal death among mothers of neonates who come from the rural area were 2.3 times higher than that of mothers of neonates who come from the urban area [

Discussion
Despite the establishment of NICU for improving the life of newborns, most of the hospital's NICU organization lacks necessary drugs and is not well equipped in which neonatal death remains to occur in those settings in Ethiopia. Therefore, we have tried to show the cause and the proportion of neonatal death in the hospital setting. Additionally, we have tried to isolate factors that in uence the poor neonatal outcome at a tertiary hospital in Ethiopia. Accordingly, we have observed that residency, LBW, RDS, sepsis, treatment with phototherapy, and oxygen were factors that signi cantly associated with poor treatment outcome in multivariable logistic regression analysis.
The present study reported that neonatal sepsis, hypothermia, and low birth weight were the predominant causes of neonatal admissions. Neonatal sepsis accounts for 55.4% of admission and 51% of neonatal death. Neonatal sepsis was found to be within the top three leading causes of neonatal admissions in several studies [13,[15][16][17]. This study implied that there is a clear gap in the management and prevention of sepsis. Similarly in this study, 49% of neonates in this study had LBW and 35.5% of death was due to LBW, which is in line with the study done in Gondar [13] and Jimma [15] and Mekele [18]. This implies that the majority of causes of neonatal admissions were attributed to preventable and treatable neonatal conditions and can be prevented if care providers practice evidence-based interventions.
In the present study, the proportion of neonatal death was inconsistent with several studies carried out in different part of Ethiopia. The current nding lower than studies conducted MizanTepi University Teaching Hospital 22.8% [12], and Gondar referral hospital, northwest Ethiopia 23.1% [19]. It also lower than studies conducted in Brazil 32.3% [20], Nigeria 20.3% [16], and Guinea-Bissau 19.6% [21]. This variation might be due to differences in socioeconomic status and quality of care provided.
The current proportion is also higher than studies conducted in India 7.16% [22], Iran 10.23% [23], Cameroon 10% [17], and Eritrea 6.6% [24]. This discrepancy might be due to the health facility setups because some of the setups could be well-equipped, with the presence of skilled manpower. Another possible reason may be in this study about 27.8% of mothers of neonates had obstetrical complications during pregnancy compared to the study done in Eritrea (6.1%) [24].
In this study mothers of neonates who came from the rural area had a 2.3 times higher risk of death. This is consistent with a study conducted in Jimma [15]. This may be due to delay in deciding to seek care and delay in reaching health care services which would affect the on-time arrival of laboring mothers to the hospitals and lead to the occurrence of obstetrical complications and then the occurrence of neonatal diseases.
In this study, the odds of death among newborns who had low birth weight were 2.5 times that of neonates who did not have the condition. This is consistent with studies done in Brazil, Guinea-Bissau, Eritrea, and Jimma, Ethiopia [15,21,24,25]. The possible justi cation is that in fact that low birth weight had immaturity of immune systems and, other body defense mechanisms that control newborn disease susceptibility [26]. Then the neonates may develop health problems including RDS, bleeding from the brain, NEC, and nally, neonatal death may follow.
Neonates having RDS had 2.86 times higher odds of death compared to those neonates without RDS.
This is consistent with studies conducted in Mexico, Ghana, and Jimma [10,13,27]. The possible justi cation is that RDS causes air leaks, intracranial bleeding, pulmonary hemorrhage, bronchopulmonary dysplasia, retinopathy of prematurity, and nally leads to neonatal death [26]. The odds of death among neonates who had neonatal sepsis were 2.48 times that of neonates who did not have the conditions. This is consistent with studies conducted in Ghana, Gondar, and Mekele, Ethiopia [10,13,18,19]. The possible justi cation is that sepsis results in abscess formation, venous thrombosis, neurologic damage, and multi-organ dysfunction [28]. Finally, neonatal mortality may follow.
Our nding in the present study also indicated that neonatal death was also related to treatment with phototherapy and oxygen. In fact that treatment with oxygen and phototherapy can cure neonatal jaundice and other neonatal illness, the decreases the chance of death [26].

Limitation of study
The limitation of this study was it might not indicate a cause-effect relationship because the study design was cross-sectional. The use of medical records of newborns because of incompleteness and since the study is institution-based, the results might lack generalization to the entire population in the catchment area.

Conclusion
The study revealed that the causes of neonatal death were neonatal sepsis, low birth weight, and prematurity. Residency, low birth weight, respiratory distress syndrome, having neonatal sepsis treated with phototherapy, and treated with oxygen were independent factors. The concerned bodies, the hospital, and health care workers should give attention to neonates admitted to intensive care units by strengthening the quality of care given at neonatal intensive care unit and strength early detection and treatment of health problems during post-partum periods. Most of the neonatal deaths are due to preventable and treatable conditions. Health care providers, hospital management should work hard to improve care for all neonates with special attention to the care of high-risk neonates and should focus on factors that affect neonatal survival to reduce neonatal mortality. Since the study was based on secondary data there was no direct contact with patients. So, anonymity was maintained by using the identi ed number instead of the patient's names. Besides, all data abstracted was kept con dential and not used for any other purpose than the stated objective and all methods were carried out per ethical guidelines.

Consent to publication
Non-applicable Availability of data and materials The data set generated or analyzed during the current study are not publicly available due to the privacy of the participants and institution restriction but are available from the corresponding author on reasonable request.

Con ict of interest
There is no con ict of interest