ADDIS ABABA UNIVERSITY

COLLEGE OF HEALTH SCEINCE

SCHOOL OF NURSING AND MIDWIFERY
DEPARTMENT OF NURSING AND MIDWIFERY

SURVIVAL STATUS AND PREDICTORS OF MORTALITY AMONG

NEONATES ADMITTED TO NEONATAL INTENSIVE CARE UNIT IN
DESSIE REFERRAL HOSPITAL, NORTH EAST ETHIOPIA, 2021. A
RETROSPECTIVE COHORT STUDY

BY: MOHAMMED KEBEDE (BSCN)

ADISOR: Dr.GIRUM SEBSIBE (MSC, PHD)
FEVEN MULUGETA (BSC, MSC)

A RESEARCH THESIS SUBMITTED TO ADDIS ABABA UNIVERSITY,

COLLEGE OF HEALTH SCIENCES, SCHOOL OF NURSING AND
MIDWIFERY, DEPARTMENT OF NURSING, FOR THE PARTIAL
FULFILMENT OF THE REQUIREMENTS FOR DEGREE OF MASTERS
OF SCIENCE IN PEDIATRICS AND CHILD HEALTH NURSING

JUNE, 2021

ADDIS ABABA, ETHIOPIA

 ADDIS ABABA UNIVERSITY
COLLEGE OF HEALTH SCIENCES
SCHOOL OF NURSING AND MIDWIFERY
POSTGRADUATE PROGRAM

Name of principal Mohammed Kebede (BSC)

investigator
Name of advisors Dr. Girum Sebsibe (MSC, PHD)
Cell phone: +251-920856732
E-mail: girumseb@gmail.com
Mrs. Feven Mulugeta (BSC, MSc )
Mobile: +251910712396
Email: fevishome@gmail.com

Full title Survival status and predictors of

mortality among neonates admitted to
neonatal intensive care unit in Dessie
Referral Hospital, Northeast Ethiopia,
2021.

Study design Retrospective cohort study

Duration From November to June 2020/2021
Study area Dessie
Total cost 26472
Address of investigator Telephone: +2519 13803674
E-mail: mohama26.mk@gmail.com
Approval sheet by the board of examiners
I, the undersigned MSc student, declare that I have submitted my original work entitled with
Survival status and predictors of mortality among neonates admitted to neonatal intensive care
unit in Dessie Referral Hospital, Northeast Ethiopia, 2021 for examination.
Submitted by:
Mohammed Kebede (BSc) __________________ ________________
Name of student Signature Date
This thesis by Mohammed Kebede is accepted in its present form by the board of
examiners as satisfying thesis for degree of masters in pediatrics and child health nursing.
Approved by
EXAMINER: _________________ ______________________
Nigusie Tadele (Assistant Professor) Signature Date
ADVISORS:
Dr. Girum Sebsibe (MSC, PHD) ________________________________
Name of major advisor Signature Date
Mrs. Feven Mulugeta (BSc, MSc) ______________________________
Name of co -Advisor Signature Date
DEPARTMENT HEAD
Name ___________________ _________________ ____________________
Nigusie Tadele (Assistant Professor) Signature Date

ii
 STATEMENT OF DECLARATION
By my signature below, I confirm that this thesis is my own novel work. I have followed all
ethical principles of research in the preparation, data collection, data analysis and
compilation of this thesis. Any scholarly matter that is included in the thesis has been given
recognition through citation. I endorse that I have cited and referenced all sources used in this
document.

This thesis is submitted in partial fulfillment of the requirements for a master degree of
pediatrics and child health nursing at the Addis Ababa University. The thesis will deposit in
Addis Ababa University library and will made available to borrowers under the rules of the
library. I seriously declare that this thesis has not been submitted to any other institution
anywhere for the award of any academic degree, diploma of certificate.

Brief references from this thesis may be made without special permission as long as
accurate and complete acknowledgement of the source is made. Requirements for permission
for extended references from or reproduction of this thesis in whole or in part may be approved
by the Head of the School or Department when in his or her judgment the proposed use of the
material is in the interest of scholarship. However, in any other situations, permission must be
obtained from the author of the thesis.

INVESTIGATOR
Name Signature Date
Mohammed Kebede ___________________ ___________________
ADVISORS:
Dr.Girum Sebsibe (MSC, PHD) ___________ _____________

Name of major advisor Signature Date

Mrs. Feven Mulugeta (BSc, MSc) _______________ __________________

Name of co -Advisor Signature Date

iii
 ACKNOWLEDGMENT
First of all, I would like to give my warmest gratitude to Addis Ababa University, College of
Health Science, School of Nursing And Midwifery, Department of Nursing and Midwifery for
giving this chance to develop this thesis. My acknowledgment also goes to Diredawa University,
for sponsoring my MSc education. I would like to express my deepest gratitude to my advisors
Dr. Girum Sebsibe (MSC, PHD) and Mrs. Feven Teshome (BSc, MSc) for their unreserved
support, provision of relevant, timely comments and guidance of the overall procedure of this
research development.

My thanks also go to Dessie Referral Hospital staffs mainly to the manager of the hospital, to
the neonatal intensive care unit case team personnel and to the card extractors. My special
thanks and appreciation goes to the supervisor and the data collectors of the study. My grateful
thanks also extend to my best friend Kirubel Bimer who support me and giving a constructive
feedback in the development of this thesis. My heartfelt gratitude also goes to my examiner
Nigusie Tadele. Last but not least my appreciation is to my all classmates, to Endris Jemal and
Rediwan Mohammed for their valuable technical support throughout my proposal and thesis
development.

iv
 ABBREVIATIONS AND ACRONYMS
AHR - Adjusted Hazard Ratio
AIDS - Acquired Immune Deficiency Syndrome
ANC - Antenatal Care
AOR - Adjusted Odds Ratio
APGAR - Appearance, Pulse, Grimace, Activity and Respiration
DM - Diabetic Miletus
EDHS - Ethiopian Demographic and Health Survey
EVLBW - Extremely Very Low Birth Weight
HIV - Human Immunodeficiency Virus
HMD – Hyaline Membrane Diseases
HSTP - Health Sector Transformation Plan
IDMs – Infant from Diabetic Mothers
MDG4 – Millennium Development Goal 4
MAS – Meconium Aspiration Syndrome
NICU - Neonatal Intensive Care Unit
NBW – Normal Birth Weight
NMR - Neonatal Mortality Rate
PNA – Prenatal Asphyxia
RDS – Respiratory Distress Syndrome
SDG – Sustainable Developmental Goal
SVD – Spontaneous Vaginal Delivery
U5MR – Under Five Mortality Rate

v
 TABLE OF CONTENTS
ACKNOWLEDGMENT ................................................................................................................... iv
ABBREVIATIONS AND ACRONYMS .......................................................................................... v
LIST OF TABLES .......................................................................................................................... viii
LIST OF FIGURES .......................................................................................................................... ix
ABSTRACT .......................................................................................................................................x
1. INTRODUCTION......................................................................................................................... 1
1.1 Background ............................................................................................................................. 1
1.2 Statement of the Problem ........................................................................................................ 3
1.3 Significance of the Study ......................................................................................................... 5
2. LITERATURE REVIEW ............................................................................................................. 6
2.1. The Magnitude of Neonatal Mortality in NICU .................................................................... 6
2.2 Predictors of Mortality in Neonates ........................................................................................ 7
2.2.1 Socio demographic factors................................................................................................ 7
2.2.2 Obstetrics and gynecologic factors ................................................................................... 8
2.2.3 Neonatal medical conditions ............................................................................................. 9
2.2.4 Maternal medical conditions .......................................................................................... 10
2.3. Conceptual Frameworks ...................................................................................................... 12
3. OBJECTIVES ............................................................................................................................. 13
3.1. General Objective ................................................................................................................. 13
3.2. Specific Objectives: .............................................................................................................. 13
4. METHODOLOGY...................................................................................................................... 14
4.1. Study Area and Period ......................................................................................................... 14
4.2. Study Design ......................................................................................................................... 14
4.3. Populations ........................................................................................................................... 14
4.3.1. Source of populations .................................................................................................... 14
4.3.2. Study population ............................................................................................................ 14
4.3.2. Sample population ......................................................................................................... 14
4.4. Eligibility Criteria ................................................................................................................ 15
4.4.1. Inclusion criteria ............................................................................................................ 15
4.4.2. Exclusion criteria ........................................................................................................... 15
4.5. Sample Size Determination and Sampling Procedure ......................................................... 15
4.5.1. Sample size determination ............................................................................................. 15

vi
 4.5.2. Sampling technique and procedures ............................................................................. 16
4.6 Study Variables ..................................................................................................................... 17
4.7 Operational Definitions ......................................................................................................... 18
4.8 Conceptual Definitions .......................................................................................................... 18
4.9 Data Collection Tools and Procedures .................................................................................. 18
4.10. Data Quality Control .......................................................................................................... 19
4.11. Data Processing and Analysis............................................................................................. 19
4.12. Ethical Considerations ....................................................................................................... 20
4.13. Dissemination of the Result ................................................................................................ 20
5. RESULTS .................................................................................................................................... 21
5.1. Socio-Demographic Characteristics of the Study Participants ........................................... 21
5.2. Obstetric and Gynecological Characteristics of the Study Participants ............................. 22
5.3 Maternal Medical Diagnosis of the Study Participants ........................................................ 23
5.4 Common Medical and Other Diagnosis of Neonates Admitted to NICU ........................... 24
5.5. Survival Status of Neonate ................................................................................................... 26
5.5.1. Death/Failure of neonates and overall Kaplan- Meier failure function ....................... 27
5.5.2. Overall Kaplan- Meier Survival Function .................................................................... 28
5.5.3. Comparison of survival time for different groups of variables .................................... 29
5.6. Predictors of Time to Death among Neonates Admitted to NICU ...................................... 39
5.7. Multicollinearity Test ........................................................................................................... 42
5.8. Test of Proportional Hazard Assumption ............................................................................ 43
5.9. Cox-Snell residuals Test ....................................................................................................... 45
6. DISCUSSION ............................................................................................................................. 46
7. LIMITATION AND STRENGTH OF THE STUDY ................................................................ 50
8. CONCLUSION ........................................................................................................................... 51
9. RECOMMENDATION .............................................................................................................. 52
REFERENCE ................................................................................................................................. 54
APPENDIX ..................................................................................................................................... 61
Appendix I Information Sheet .................................................................................................... 61
Appendix II Checklist ................................................................................................................. 62
Appendix III Approval Sheet...................................................................................................... 66

vii
 LIST OF TABLES
Table 1: Sample size calculation to assess the survival status and predictors of mortality among
neonates admitted to NICU in Dessie Referral Hospital, Northeast Ethiopia, 2021 ................ 16
Table 2: Socio-demographic characteristics of neonate and their mother in Dessie Referral
Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020. .............................. 21
Table 3: Obstetric and gynecological characteristics of the study participant in NICU of Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020. ................. 22
Table 4: Maternal medical diagnosis in NICU of Dessie Referral Hospital, Northeast Ethiopia,
January 1st, 2018 – December 31th, 2020. ............................................................................ 23
Table 5: Common medical and other diagnoses of neonates that were admitted to NICU of
Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020. ...... 24
Table 6: Median survival time and log-rank test for equality of survivor functions among
neonate admitted to NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 –
December 31th, 2020. ........................................................................................................... 37
Table 7: Results of bivariate and multivariate analysis using Cox regression model for predictors
of time to death among neonate admitted to NICU, in Dessie Referral Hospital, Northeast
Ethiopia, January 1 st, 2018 – December 31th, 2020. ............................................................... 40
Table 9: Multicollinearity check for each variable based on binary outcome (p-value<0.25) to
neonates admitted in NICU at Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018
– December 31th, 2020 ......................................................................................................... 42
Table 8: Schoenfeld Residuals test for Cox-Proportional Hazard Regression model assumption
(estat, phtest) of each variable for neonates admitted in NICU at Dessie Referral Hospital,
Northeast Ethiopia, January 1st, 2018 – December 31th, 2020. ............................................. 43
Table 10: Data collection checklist for conducting of study on survival status and predictors of
mortality among neonates admitted to NICU in Dessie Referral Hospital, Northeast Ethiopia,
2020...................................................................................................................................... 63

viii
 LIST OF FIGURES
Figure 1: Conceptual framework for the survival status and predictors of mortality among
neonates admitted to NICU in Dessie Referral Hospital, Northeast Ethiopia, 2021 ................ 12
Figure 2: Schematic presentation of sampling procedure to assess the survival status and
predictors of neonatal mortality among neonates admitted to NICU in Dessie Referral Hospital,
Northeast Ethiopia, 2021 ....................................................................................................... 17
Figure: 3 Outcome of neonates admitted to NICU of Dessie Referral Hospital, Northeast
Ethiopia, January 1st, 2018 – December 31th, 2020. ............................................................. 26
Figure 4: Overall Kaplan-Meier failure estimate of neonate admitted to NICU of Dessie Referral
Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020. .............................. 28
Figure 5: Overall Kaplan-Meier survival estimate of neonate admitted to NICU of Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020. ................. 29
Figure 6: The Kaplan-Meier survival curves compare survival time of neonate with groups of
prolonged labor at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 –
December 31th, 2020 ............................................................................................................. 30
Figure 7: The Kaplan-Meier survival curves compare survival time of neonate with groups of
RDS at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December
31th, 2020. ............................................................................................................................ 31
Figure 8: The Kaplan-Meier survival curves compare survival time of neonate with groups of
sepsis at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December
31th, 2020. ............................................................................................................................ 32
Figure 9: The Kaplan-Meier survival curves compare survival time of neonate with groups of
anemia at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December
31th, 2020. ............................................................................................................................ 33
Figure 10: The Kaplan-Meier survival curves compare survival time of neonate with groups of
bag and mask resuscitation at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January
1st, 2018 – December 31th, 2020. ......................................................................................... 34
Figure 11: The Kaplan-Meier survival curves compare survival time of neonate with groups of
first minute APGAR score at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January
1st, 2018 – December 31th, 2020 .......................................................................................... 35
Figure 12: The Kaplan-Meier failure curves compare survival time of neonate with groups of
fifth minute APGAR score at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January
1st, 2018 – December 31th, 2020 .......................................................................................... 36
Fig 13: Cox-Snell residual cumulative hazard graph on neonates admitted in NICU at Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020 .................. 45

ix
 ABSTRACT
Background: The first 28-days of life, particularly the day of birth, are the most challenging
and vulnerable time for a child’s survival and health. Neonatal deaths as a share of under-five
deaths decreased more slowly than mortality among children aged 1–59 months and accounted
for 47 percent of all under-five deaths. Moreover, Ethiopia is still on the 4th and 2nd rank of
global and sub-Saharan Africa neonatal mortality rate, respectively, which indicates neonatal
death is a critical problem which needs farther studying of the underline predictors. Even though,
some studies were conducted on predictors and determinants of neonatal mortality in Ethiopia,
little is known about the time to neonatal death and its predictors. Therefore, this study aimed
to assess the survival status and predictors of mortality among neonates admitted to NICU in
Dessie Referral Hospital. Objectives: To determine the survival status and predictors of
mortality among neonates admitted to NICU from 01/01/2018-31/12/2020 in Dessie Referral
Hospital, Northeast Ethiopia, 2021. Methods: An institution-based retrospective cohort study
was used among 542 neonates admitted from 2018-2020 in NICU of Dessie Referral Hospital,
Northeast Ethiopia. Medical record numbers were obtained from patient register logbooks. Then
data was collected from patient cards by using a systematic sampling method with pretested
checklist. Data was entered using Epi-data 4.6 and analyzed using STATA 16. Kaplan Meier
was used to estimating median survival time and cumulative probability of survival, and the log-
rank test was used to compare survival curves. The Cox proportional hazard regression model
was used to analyze the relationship between independent and outcome variables. Results: Out
of 542 eligible participant neonates, 104 (19.19%) died with an overall incidence rate of 35.54
(95%CI: 29.33, 40.08) deaths/1000 neonate-days observations. The overall median survival
time was 14 days, with an overall cumulative survival probability of 36.10%. Low birth weight
(AHR: 3.03, 95%CI: 1.53, 6.00), prolonged labor (AHR: 3.32, 95%CI: 1.46,7.53), primiparous
mother (AHR: 2.23, 95%CI: 1.44, 3.45), preeclampsia (AHR: 2.17, 95%CI: 1.10, 4.25),
Diabetic mother (AHR: 3.74, 95%CI: 1.57, 8.90), PNA (AHR: 2.61, 95%CI: 1.11, 6.13), and
neonatal anemia (AHR: 9.14, 95%CI: 3.33, 25.08) were independent predictors of death.
Conclusion: The incidence of death was 35.54 per 1000 neonate-day and LBW, primipara
mother, prolonged labor, preeclampsia, maternal DM, PNA, neonatal anemia were identified as
independent predictors of time to death of neonates. Keywords: Neonate, Survival Status,
Neonatal Mortality, Predictors, Time to Death

x
 1. INTRODUCTION
1.1 Background
In the first 28 days of life, infants are known as neonates and are the most critical age for child
survival. The death of infants within the 28 days of life is called neonatal mortality and is stated
as neonatal deaths/1000 live births. It is estimated that worldwide, 2.5 million neonatal death
occur each year, contributing 47% of under-five mortality. Among deaths, close to three-
quarters of all neonatal deaths occur in the first week after birth, and 36% die on the day they
are born (1).

Globally, an estimated 19 million newborn babies suffer from preterm birth, intrapartum-related
brain insults, severe bacterial infection, and pathological jaundice, which causes mortality and
survival with long-term complications every year (2). In developing countries, prematurity
(39.8%), perinatal asphyxia (20.8%), neonatal sepsis (17.1%), and congenital malformation
(15.7%) were the major cause of neonatal deaths (3).

In sub-Saharan Africa, infections, birth asphyxia, complications of preterm birth, and low birth
weight are the most common cause of neonatal mortality. Five countries shared 50% of neonatal
death in this region; Nigeria, the Democratic Republic of Congo, Ethiopia, Tanzania, and
Uganda (4). Moreover, in East Africa, a study showed that home births, mothers without an
education and mothers whose husbands decided on contraceptive practices, and rural residency
are predictors of neonatal mortality (5).

Between 1990 and 2017, the estimated global neonatal death fell by half from 5 million to 2.5
million and the mortality rate decreased from 37 deaths/1,000 live births to 18 (6). Despite
decreasing neonatal mortality globally, marked differences exist across regions and countries.
For example, in 2017, the highest neonatal mortality was found in sub-Saharan Africa, with 27
deaths/1,000 live births. Neonates born in this region had 9 times the risk of death in the first
month than neonates in high-income countries. Likewise, the neonatal mortality rate ranged
from 1 death per 1,000 live births to 44 deaths across countries. (1).

Ethiopia is the sixth among ten countries with highest neonatal death numbers with 87800 deaths
annually and showed slow progress in reducing neonatal mortality rate (NMR) (2). According
to the 2019 Ethiopia Mini Demographic and Health Survey (2019 EMDHS) report, Ethiopia's

1
neonatal mortality rate is 30 per 1000 live birth (7). Moreover, a study done in Amhara Regional
State Referral Hospitals investigated that NMR per 1000 live birth was 27 (8).

In Ethiopia, studies showed that parental residency, the length of stay in NICU, low birth weight,
prematurity, respiratory distress syndrome (RDS), perinatal asphyxia (PNA), congenital
malformations, meconium aspiration syndrome (MAS), low 5min Apgar score, time of initiation
of breastfeeding, and hyperthermia, were predictors of neonatal mortality in the neonatal
intensive care unit (NICU) (9,10).

2
1.2 Statement of the Problem
The first 28-days of life, particularly the day of birth, are the most challenging and vulnerable
time for a child’s survival and health. Every day, an approximated 7,300 newborns die globally
from complications of prematurity, intrapartum related deaths, and neonatal infections, and most
of these causes are preventable and treatable (11). Worldwide around 2.4 million neonates died
in 2019, with the highest neonatal mortality rate in sub-Saharan Africa and South Asia (12,13).
Among those deaths, 99000 were from Ethiopia, which holds the 4th rank globally next to India,
Nigeria, and Pakistan (13).

According to Ethiopia demographic and health survey (EDHS 2016) report, in Ethiopia, 1 in 35
children dies in the first month of life. This report also found that shorter intervals between
births (< 2 years gap between birth), being a boy newborn, being small or very small newborn
at birth, born from women age 34 or older, born from women of age less than 18, and born from
mother of high parity (birth order higher than three) were the contributed factors of those deaths
(14).

Despite the declining global child mortality rate, Sub-Saharan Africa remains with the highest
under-five mortality rate (U5MR), and 39 percent of global newborn deaths occurred in this
region (1,15). Two-thirds of neonatal mortality is reported from 12 countries, six of which are
in sub-Saharan Africa with a rate of 30 or more deaths per 1 000 live birth (16). Also, there is a
difference in neonatal survival between the poorest and the wealthiest regions and countries. In
sub-Saharan Africa (1 in 37) and in Europe and Northern America (1 in 333) child die in the
neonatal period (1). Moreover, neonatal deaths as a share of under-five deaths decreased more
slowly than mortality among children aged 1–59 months and accounted for 47 percent of all
under-five deaths (1,17).

Ethiopia has well achieved its target for MDG 4, reducing child mortality by two-thirds, before
three years to the target year 2015; however, its achievement in reducing neonatal mortality
remains unchanged (30 per 1000 live births) (7,18) and is still on the 4th and 2nd rank of global
and sub-Saharan Africa neonatal mortality rate, respectively, which indicates neonatal death is
a critical problem which needs farther studying of the underline predictors (13). The Ethiopian
government showed its effort to improve neonate’s survival by including pronounced life-saving
interventions in its national strategy for Newborn and Child Survival and Health Sector

3
Transformation Plan (HSTP) (19,20). Regardless of these efforts, neonatal mortality is still high
and needs further investigation of the root causes.

In the era of Sustainable developmental goal (SDG), decreasing neonatal mortality is an integral
part of goal 3 to end preventable deaths of newborns and under-five children to reduce neonatal
death to at least 12/1,000 live births by 2030 (1,6,11,17). However, with the current neonatal
death rate trend, Ethiopia needs to accelerate at least 2 times the current reduction rate of
progress to meet the NMR targeted by SDG 2030 (6,17). Therefore investigating the predictors
of time to neonatal death is the key to attain the goal.

In Ethiopia, some studies were conducted on the predictors and determinants of neonatal
mortality. However, little is known about the time to deaths of neonates and its predictors. This
study will also find the median survival time of neonates and their predictors. Moreover, no
published study is found on a similar topic in the study area as my literature review. Therefore
this study is important to show the incidence density rate of neonatal mortality and its predictors
in the study area. Furthermore, maternal medical condition predictors are added in detail in this
study as compared to other studies conducted in Ethiopia. Currently, reducing neonatal mortality
is the global as well as an Ethiopian priority health issue. Therefore, this study aimed to assess
the survival status and predictors of mortality among neonates admitted to NICU in Dessie
Referral Hospital.

4
1.3 Significance of the Study
At this time, studying of neonatal mortality and its predictors is very important because,
reducing neonatal mortality is a prioritized issue in the world as well as in Ethiopia as part of
the third Sustainable Development Goal (SDG), to end preventable child deaths. Therefore to
achieve this goal, an understanding of the predictors of neonatal mortality will require (6).

This study will contribute to the implementation of global and national intended strategies and
programs towards reducing NMR by providing data that is essential in forwarding planning,
especially budgeting and facilities, staffing, and training with the aim of improving the outcome
of neonates in the healthcare settings.

This study will use as baseline-information for future researchers who will be interested to do
their research on related topics. It will also useful for health policymakers to design appropriate
policies, strategies, and interventions.

Findings from this study in the study area will also offer input to decision-makers, program
implementers, monitors, and evaluators, to advance better maternal and newborn care. It also
improves mothers’ awareness about predictors of neonatal mortality and the importance of early
care-seeking behavior and birth preparedness.

The study results will be useful for health professionals in identifying predictors of time to death
of neonates and to take intervention accordingly. This study will also increase the nursing
knowledge and nursing practice, and promote nursing education and research.

5
 2. LITERATURE REVIEW
2.1. The Magnitude of Neonatal Mortality in NICU
Worldwide, an estimated 2.5 million newborns, approximately 7,000 every day, died in the first
28 days of life in 2017 (1). Among these deaths, 98% occur in low-resource countries, including
Ethiopia (21). Thus, around three-quarters of all newborn deaths is early neonatal death in the
first week of life. In addition, the shared neonatal mortality among under-five deaths increased
from 40 percent in 1990 to 47 percent in 2017 (1), indicating that currently, around half of
under-five death occurred in the neonatal period.

A systematic and meta-analysis study in the NICU of Iran showed that 11.4 percent of neonates
died (22). Another Hospital-based prospective and retrospective study in India revealed that the
overall neonatal mortality rate was 9.73 percent (23) and 10.4 percent (24). UNICEF’s report
in 2015 showed that the incidence of neonatal mortality in Afghanistan was 36 per 1000 live
birth (25). Different studies in Africa reported that neonatal mortality rate was 22% in rural
Ugandan NICU (26), 14.2% in Nigeria Enugu State University Teaching Hospital (27), 10.7 %
in Northern Tanzanian neonatal care unit (NCU), in which 56.7% of those deaths occurred in
the first 24 hours, and 91.6% within the first week of life (28). Similarly, a study done in
Mauritania revealed that 34.7% of neonates died, and among those deaths, 71.3% were during
the first six days of life (29).The incidence of NM in the Democratic Republic of Congo was 94
per 1000 live births (30).

In Ethiopia, the neonatal mortality rate has fallen from 39 to 29 deaths /1,000 births between
2015 and 2016 but has remained constant since 2016-2019 (7). A study conducted in NICU at
Arba Minch General Hospital discovered that the IDR of neonatal mortality was 31.6/1,000
neonate days. The 1st day of survival probability was 96.1%. This study also showed that
survival probability at the 7th, 14th, and 21st day was 75%, 69.9%, and 66.2%, respectively, with
an overall median survival time of 6 days (31). Also, a study in Jimma University Medical center
showed that the incidence of NM was 30 deaths per 1,000 institutional live births (10) and 27
per 1000 neonates-days (32). A study done in the University of Gondar Comprehensive
Specialized Hospital (UoGCSH) revealed that 14% of the neonatal death was in the first 24 hrs,
60% in the first 3 days, and almost 84% in the first 1 week of life. Also, the failure probability
was 13.79%, 59.77%, 83.91%, and 97.70% in the 1st, 3rd, 7th, and 14th days of hospital stay

6
with an incidence rate of 23.81, 33.92, 31.48, and 28.90 deaths per 1000 neonate-days
observation, respectively (33). Another study conducted in the Tigray region revealed that the
neonatal mortality incidence was 62.5 per 1000 live birth. Similarly, studies done in the NICU
of Arba Minch General Hospital and Debre Markos Referral Hospital discovered 20.8% and
21.3% deaths, respectively (31,34). The failure probability at the 1st, 3rd, and 7th, days of
hospital stay was 19.4%, 58.2%, and 91%, respectively, and there was no failure after 16
days(31). All the above findings indicated that NM incidence in Ethiopia is far from the united
nations SDG targeted to achieve 12 or fewer neonatal death per 1000 live birth by2030 (35).

Studies done in Ethiopia showed that survival time significantly differed among the covariates
of neonatal age at admission, 1st and 5th minute APGAR score, breastfeeding initiation, place
of delivery, prolonged Labor, residence, PROM, and birth weight with log-rank test (p-
value<0.05). However, there was no survival time difference between categories of sex,
maternal age, ANC follow-up, mode of delivery, GA, number of parity, and multiple
pregnancies with a p-value >0.05 (31,36).

2.2 Predictors of Mortality in Neonates

2.2.1 Socio demographic factors
Different findings revealed that low birth weight, gestational age, maternal age at birth, and
residence were the most common Socio-demographic predictors of neonatal mortality
(24,28,31,37–40). For example, studies showed that neonatal survival was significantly
associated with low birth weight (< 2500 gm) (9,24,29,37,38,40–44).

A study in the Democratic Republic of Congo revealed that neonates born from mothers age
less than 18 and more than 35 years had a high risk of death (OR=9.65) (45). A meta-analysis
study showed that Nulliparous/age <18 years (AOR: 2.07) (46). In Ethiopia, newborns from
mothers less than or equal to18 years old accounts for most of the neonatal death compared to
babies born from mothers aged 35 years or above (AHR: 12.6) (47). Another study showed that
neonates born from mothers of advanced age were a higher risk of death (AHR=2.60) (8).

Regarding gestational age, the risk of death in preterm (before 34) was 4.5% higher than babies
born after 34 weeks, and the risk becomes higher at 32 weeks (10.17 times) (P<0.0001) (24).
In another study, the deaths of preterm babies were higher (24.5%) than deaths of term babies
(7.1%) (38). Research done in South Africa reported that being preterm delivery (OR=2.975)
7
was a significant predictor of neonatal mortality (43). Similarly, a study showed that neonatal
mortality tends to occur with gestational age lower than 37 weeks (p = 0.000) than gestational
age greater than 37 weeks (37). In other findings being preterm was significantly associated
with increasing time to neonatal death (Pooled Hazard Ratio (PHR): 3.78) (48).

A study conducted in sub-Saharan Africa revealed that rural residence was a significant
predictor of time to neonatal death (PHR: 1.91) (48). Neonatal death was high both in the early
and late neonatal periods in rural areas than in urban area (44). Research done in South Africa
reported that being male babies (OR=1.633) was a significant predictor of NM (43). But
research conducted in Nigeria showed that gender was not a significant predictor of neonatal
mortality (p > 0.05) (37).

2.2.2 Obstetrics and gynecologic factors

Different studies in different countries discovered many causes of neonatal mortality and
survival and obstetrics and gynecology predictors were among the major predictor that
contribute to neonatal death. A study conducted in Northwestern rural China on predictors of
neonatal death revealed that neonatal death was higher among multiparous women (OR= 2.77)
(49). A Study done in Sub-Saharan Africa revealed that multiple births (being twins or triplets
or more) was independent predictors of time to neonatal death (48). Similarly, studies in
Ethiopia showed that multiple pregnancies increased neonatal death risk as compared to a single
pregnancy with (AHR=3.96) (8,32).

Concerning Antenatal care (ANC), NMR was higher among mothers who did not receive ANC
in the first trimester of pregnancy (OR= 2.49) (49). A Study in the Democratic Republic of
Congo also showed that being born from mothers with ANC follow-up <4 was a predictor of
NM (OR=9.59) (45). A study in NICU at Referral Hospital in Southern Ethiopia discovered that
having no ANC visits during pregnancy significantly increases neonatal deaths with AHR=6.02
(32) and a study done in NICU at Debre Markos Hospital with AHR: 1.9 (34). Research done
in Assosa zone and Ayder comprehensive hospital (Mekele) revealed that neonatal mortality
was high among primipara mothers neonate (50,51). Another study showed that prolonged
rupture of the membrane increases the risk of time to neonatal death with AHR:2.6 (31).

A study from middle-income countries reported that cesarean section was a significant predictor
of NM with OR = 1.8 (42). Similarly, a study done in Afar Region Public Hospitals found that

8
NM was high among those delivered by cesarean section than those delivered by spontaneous
vaginal delivery (SVD) with AOR=3.52 (52). However, Southern Ethiopia research revealed a
66% protective effect on neonatal death compared with SVD (32). Besides, a study done in
NICU at the University of Gondar Hospital showed that instrumental delivery had a high risk
of neonatal mortality with AOR: 2.99 (53). Also, another study showed that NM was higher in
assisted delivery compared to SVD (AOR=3.28, P<0.05) (52).

A study conducted in Afghanistan investigated that ANC by a skilled provider (AOR: 0.7), and
institutional delivery (AOR: 0.7), had reduced NM (54). Research done in East and West Africa
( Mozambique (OR = 1.4), and Niger (OR=2.6),) found that home delivery had a higher risk of
neonatal mortality (5). A study done in the NICU of the UoGCSH showed that home delivery
(AHR: 2.43) was a statistically significant predictor of time to death of neonates at a p-value ≤
0.05 (33).

2.2.3 Neonatal medical conditions

As investigated by different studies, neonatal medical problems are the primary cause of
neonatal mortality. For example, a systematic Review and Meta-Analysis study done in Iran on
neonatal mortality in NICU found that (RDS) (31.93%), septicemia (12.66%), and asphyxia
(7.58%) were the most common predictors of neonatal mortality (22). Other studies in India and
Bangladesh also revealed that RDS (33.6%) (24), PNA (25.3% and 24.3%), and sepsis (25.3%
and 21.9%) were predictors of neonatal death, respectively (24,40). Also, research done in
Vietnam found that infection (38%), cardio/respiratory disorders (27%), and congenital
abnormalities (20%) were the main causes of neonatal deaths (55).

Research done in sub-Sahara Africa discovered that sepsis (PHR: 1.94), PNA (PHR: 2.03), and
RDS (PHR: 2.49) were independent predictors of time to neonatal death (48). PNA was an
independent predictor of mortality (AOR =5.817) in the study done in tertiary care hospitals of
Addis Ababa (56). Likewise, a study done in Dessie Referral Hospital on determinants of NM
discovered that RDS (AOR = 3.61), PNA (AOR = 2.27), MAS (AOR = 2.35), and infection
(AOR = 2.26) were significantly associated with NM (57).

Low first and 5th minute Agar score (PHR: 3.47) and neonates who did not initiate breastfeeding
within one hour of birth (PHR: 3.77) were independent predictors of time to neonatal death in
the study found in sub-Sahara Africa (48). Moreover, neonates for whom breastfeeding was not

9
initiated within the first 24hours were also more likely to die (AOR=12.16) (9). Studies
conducted in public General hospitals of Eastern Ethiopia and Arba Minch General Hospital
revealed that the fifth minute APGAR scores ≤5 had a higher risk of neonatal death with AHR:
5.2 (9,47).

A study done in NICU of UoGCSH showed that neonates born before 34 weeks had a higher
hazard of death (AHR: 3.25) (33). Furthermore, resuscitated neonates had a higher hazard of
death than the counterparts with AHR 2.28 (32). Studies showed that fever on admission
(AOR=6.68) (9) and hypothermia (<35.5°c) on admission (AHR= 1.58) (32) were significantly
associated with time to death of neonates.

2.2.4 Maternal medical conditions

Different studies discovered that neonates born to mothers with the medical problem had a low
survival rate. In addition, neonatal mortality was higher among neonates born from mothers
with a history of illness than neonates born from mothers without a history of illness during
pregnancy.

A study in Mozambique found that the prevalence of maternal anemia (49.4%), severe anemia
(4.1%) at p<0.001, and syphilis (6.6%) were higher in HIV- infected women than in those HIV-
uninfected women (40.6% and 1.8%, p = 0.004) and (1.8%, p<0.001). Likewise, new born
anemia was higher among neonate born from HIV-infected mothers (10.6% vs 7.3%, p = 0.022)
(58). A study done in Amhara Region revealed that the risk of death of neonates born from
HIV-positive mothers was higher than those neonates born to HIV-negative mothers (AHR
=6.57) (8).

Regarding maternal hypertension and neonatal outcome, a study conducted on 60 mothers with
their baby revealed that 73.33% of babies were born as preterm, 76.66% were born as low birth
weight, 50% had respiratory distress, 13.33% had PNA with seizure, 20% had meconium
aspiration syndrome (MAS), and 38.33% had suspected sepsis (59). All these conditions are the
most predictors of neonatal mortality. A study done in Rwanda found that babies born from
hypertensive mothers had low birth weight (75.4%), were preterm (59.6%), had intrauterine
growth restriction (32.4%).

10
Research done in India on neonatal outcome in infants born to diabetic mothers found that their
neonates had hypoglycemia (54%),RDS (16%), Congenital anomalies (40%), out of which
cardiac anomalies (85%) (60). All of these problems are the most predictors of neonatal
morbidity and mortality. A study conducted in Nigeria found that hypoglycemia (63.8%) and
hyperbilirubinemia (57.4%) were significantly higher in IDMs than non-IDMs (61).

11
2.3. Conceptual Frameworks
After reviewing different literatures (5, 9, 21, 23, 29-31, 33, 38, 40, 42, 44-51) about predictors
of neonatal mortality, the following conceptual frame work is adapted and showing the
interaction between sociodemographic, maternal medical condition, obstetrics and
gynecological, neonatal medical predictors, and institutional factors and association between
dependent and independent variables.

Sociodemographic predictors
• Age of the mother
• Gestational age
• Sex of the baby
Obstetrics and • Residence
Gynecological • Birth weight
predictors
• Mode of delivery
• Place of delivery
• Parity Maternal medical
. • Prolonged labor condition
…… • PROM Time to death of predictors
• History of ANC neonate
follow-up • HIV/AIDS
• Type of • Hypertensi
pregnancy Neonatal medical condition on
(single or
predictors • DM
multiple) • Anemia
• Preeclampsia • PNA and MAS • HBV
• Placenta previa • RDS • CHF
• Sepsis • Epilepsy
• Hypoglycemia
• Neonatal jaundice
• APGAR score
• Resuscitation at birth
• Time of initiation of BF
• Temperature at
admission
• Neonatal anemia

Figure 1: Conceptual framework for the survival status and predictors of mortality among
neonates admitted to NICU in Dessie Referral Hospital, Northeast Ethiopia, 2021

12
 3. OBJECTIVES
3.1. General Objective
To determine survival status and predictors of mortality among neonates admitted to NICU from
01/01/2018-31/12/2020 in Dessie Referral Hospital, Northeast Ethiopia, 2021.

3.2. Specific Objectives:

To determine survival status of neonates admitted to NICU from 2018-2020 in Dessie Referral
Hospital, Northeast Ethiopia, 2021.

To assess survival time difference between groups of variables among admitted neonates to
NICU from 2018-2020 in Dessie Referral Hospital, Northeast Ethiopia, 2021.

To identify predictors of time to death among admitted neonates in NICU from 2018-2020 in
Dessie Referral Hospital, Northeast Ethiopia, 2021.

13
 4. METHODOLOGY
4.1. Study Area and Period
The study was carried out in Dessie Town, a South Wollo Zone major city under Ethiopian’s
Amhara Regional State. Dessie Town is located 401km away from Addis Ababa, the capital city
of Ethiopia, and 480 km from Bahir Dar, the capital city of Amhara Regional State. It has one
governmental and three nongovernmental hospitals. Dessie referral Hospital is the only
governmental Hospital with 597 beds in medical, gynecological and obstetrics, surgical,
pediatrics, emergency and outpatient department (OPD), and NICU. It serves approximately
7,000,000 population and has 246 nurses, 12 Anesthesiologists, 103 General Physician (GP), 24
surgeons, 34 midwives, and 48 laboratory technicians.

The NICU of Dessie Referral Hospital ward has a maximum of 35 beds with average of 6
patient’s daily admission. There are on average 1900-2200 annual admissions of neonates and
averagely 32% of admissions are from referrals of other birth centers.

The study was conducted from February 08-March 08/ 2021.

4.2. Study Design

An institution-based retrospective cohort study was conducted among neonates admitted at
NICU at Dessie Referral Hospital, Northeast Ethiopia.

4.3. Populations
4.3.1. Source of populations
This study’s source of populations were all neonates admitted to NICU at Dessie Referral
Hospital.

4.3.2. Study population

The study populations were all selected neonates admitted at NICU from the first of January
2018- the end of December 2020 at Dessie Referral Hospital.

4.3.2. Sample population

All neonates admitted to NICU at Dessie Referral Hospital, from January 2018- December 2020
and neonates who fulfilled the inclusion criteria.

14
4.4. Eligibility Criteria
4.4.1. Inclusion criteria
All admitted neonates during the study period (01/01/2018-31/12/2020) in NICU at Dessie
referral hospital.

4.4.2. Exclusion criteria

Incomplete records of registered neonates, lost record of registered neonates, and registered
neonates with an outcome beyond the 28 days of age during the follow-up period.

4.5. Sample Size Determination and Sampling Procedure

4.5.1. Sample size determination
A single population proportion formula was used to estimate the sample size for the survival
status (outcome) of neonates, by using:

P = the proportion of neonatal mortality among neonates admitted in NICU, 21.3% from a study
done in Debre Markos, Ethiopia (34)

Z α/2 = the resultant Z score of 95% CI and α=0.05

d= a tolerable margin of error (5%) and N= the minimum sample size required

(𝑍𝛼/2)2×𝑃(1−𝑃)
N= (𝑑)2

N= (1.96)2 × 0.213×0.787/ (0.05)2 = 258. After taking a 10 % non-response rate, the final sample
size will be 284.

For predictors of neonatal mortality, the double population proportion formula was used to
estimate the sample size by using the weight of neonate, RDS, congenital malformation,
prematurity, and multiple pregnancies as the major predictor variables. Then, a variable gives
the largest sample is used as the independent predictor, which is prematurity (to decrease the
role of chance) (Table1). The sample size is calculated by using Epi info version 7.2.3.1
statistical package.

𝑍𝛼 1 𝑃2(1 − 𝑃2)
[ √(1 + ) 𝑃1(1 − 𝑃1) + 𝑍𝛽√𝑃1(1 − 𝑃1) + ]2
2 𝑟 𝑟
(𝑃1 − 𝑃2)2

15
Table 1: Sample size calculation to assess the survival status and predictors of mortality
among neonates admitted to NICU in Dessie Referral Hospital, Northeast Ethiopia, 2021

Variable Assumption Total sample size After adding 10%

non - response rate
Weight of neonate P1=0.312 114 (32) 125
(low birth weight) P2= 0.087
RDS P1= 0.40 130 (34) 143
P2= 0.165
Congenital P1= 0.06 198 (32) 218
malformation P2= 0.204
Prematurity P1= 0.286 536 (34) 590
P2= 0.18
Multiple birth P1= 0.48 60 (32) 66
P2= 0.12

Where:
P1 = is the percent of exposed with the outcome
P2 = is the percent of non-exposed with the outcome
Z α/2 = is taking 95% CI, which is 1.96.
ZB = 80% power
And r is the non-exposed to the exposed ratio (1:1)
Finally, the sample size selected for this study was 590.
4.5.2. Sampling technique and procedures
There were 6336 patients admitted to NICU from January 2018-December 2020 when obtaining
the number of admission for each year. The medical record number (MRN) of the neonate was
used as a sampling frame. The sample was proportionally allocated for each year. Study
participants were selected using a systematic random sampling method until obtained the
required sample size during the actual data collection period. The study participant neonates of
each year were chosen as follows. First, determine the sampling interval (K) value by dividing
the total admitted neonate in the study period by the total sample size, which gives 10.74 ≈ 11.
Then, each year’s desired sample size was calculated by dividing the number of admission
within each year by the sampling interval (K). Finally, one number was selected between 1 and
11 at random to start sample selection, and the number is included in the sample (in this case, 9
was selected). Then, the next selection was continued every 11th unit after 9th.

16
 Total number of neonate (2018-2020) = 6336

(2018) (2019) (2020)

N=1944 N=2112 N=2280

Proportional allocation for each year

year

n= 181 n= 212
n=197

Using Systematic sampling method for each year

Final sample size = 590

Figure 2: Schematic presentation of sampling procedure to assess the survival status and
predictors of neonatal mortality among neonates admitted to NICU in Dessie Referral Hospital,
Northeast Ethiopia, 2021

4.6 Study Variables

4.6.1 Dependent variable
Time to death of neonate

4.6.2. Independent variable

Sociodemographic factors: Birth weight, Sex, GA, Maternal age, Residence

Obstetrics and gynecological factors: Mode of delivery, delivery attendant, Place of delivery,
Parity, PROM, prolonged labor, preeclampsia, placenta previa, ANC, Multiple Pregnancy, and
Place of delivery

Neonatal medical condition factors: PNA, MAS, RDS, HMD, Sepsis, Hypoglycemia, Low
Apgar score, having Resuscitation, Temperature at admission, Time of initiation of BF,
Neonatal anemia

17
Maternal medical condition factors: HIV/AIDS, Hypertension, DM, Anemia, HBV, CHF,
Epilepsy

4.7 Operational Definitions

Censored: The neonate’s outcome during the study period discharge to home (recovered),
discharged against medical advice, or transfer out to other health institutions with an unknown
outcome.

Follow-up time- The time from admission until either death or censorship occurs.

Survival status: outcome of the neonate; either death or censored

Survival time: Measures the follow-up of time from admission to the occurrence of the outcome
(death or censorship).

4.8 Conceptual Definitions

Early neonatal period: The baby considered that after birth, the first 7days of life

Having ANC follow up- Mother who has a follow up of ≥ 4 times during her pregnancy.

Medical disorders in the mother: Any registered medical diagnosis history of the mother on
the neonate’s medical record.

Multiparty: Documented maternal parity of >2 on the neonate’s medical record.

Late neonatal period: Neonate after 7 days of life after birth

Low APGAR SCORE: A neonate who have an Apgar score of <7.

Preterm (premature) birth: Neonate born before 37 weeks of gestational age.

Premature rupture of membrane: the rapture of membrane 12 hrs before the onset of labor

Prolonged labor: labor continuing beyond 18 hrs without delivering the baby

4.9 Data Collection Tools and Procedures

The tools were adapted from different literature and modified according to the Ethiopian context
(8,10,32–34,44,48). The questionnaire was prepared as a checklist in the English language. The
tool consists of sociodemographic characteristics neonates and mothers, maternal and neonatal
medical disorders, gynecologic and obstetric factors, date of admission, date of discharge, and
18
outcome of the neonate. Two BSC nurse data collectors retrospectively reviewed the neonate’s
medical records that fulfill the inclusion criteria admitted to NICU at Dessie Referral Hospital
during the study period. The survival status of patients and predictors were obtained from the
medical record. Survival time was obtained by considering the time between the dates of
admission to date of discharge (the date of death or censored).

4.10. Data Quality Control

The checklist was pre-tested among 5% of the study participants in the study area (Dessie
Referral Hospital) to assure the data quality. The necessary modifications and corrections were
made to standardize and ensure its validity. The data collectors and supervisor were senior
experienced BSc nurses who are working in NICU (one supervisor and two data collectors were
selected). One day, training was given on the purpose of the study, the data collection tool
(checklist), data collection methods, and ethical concerns during data collection. The supervisor
had monitored the data collection process of the data collectors. The Principal investigator and
supervisor had done spot-checking and reviewing the complete checklist by the data collectors
to ensure the completeness and consistency of the collected information (data). Consistency was
examined through a random selection of cards by the principal investigator.

4.11. Data Processing and Analysis

Data were entered using Epi-Data version 4.6.0 and analyzed by STATA 16 statistical software.
Data were cleaned and edited before analysis. Neonates' cohort characteristics of continuous
data were described by mean and standard deviation. Categorical data were calculated using
frequency distribution. Then, the outcome of each participant neonate was dichotomized into
censored or death. Incidence density rate (IDR) was estimated for the entire study period and
specified day intervals of hospital stay. Consequently, the NMR within the study period was
divided by the total neonate day at hazard on follow-up and reported per 1000 neonate day.
Kaplan Meier was used to calculating median survival time cumulative probability of survival.
The log-rank test was used to compare survival curves after admission to NICU. Cumulative
survival probability at certain time intervals was estimated using a life table. The Cox
proportional hazard regression model was used to analyze the relationship between independent
and outcome variables. Bivariate analysis was done for all variables and, variables with a p-
value ≤ 0.25 (62) were transferred to multivariable analysis. Before multivariate analysis,

19
multicollinearity was tested using the VIF command to check variable inflation factor (mean
VIF=1.62) (table 9). Then proportional hazard assumptions were checked using both Schoenfeld
residual test (global test) with a value of P>chi2= 0.4655 (table 8) and concordance probability
test (Harrell’s c= 0.8286), which means that the predictive ability of the survival time model for
pair of data was 82.86%. Also, Cox proportional hazard model fitness to the data was cheeked
using cox Snell residuals, in which the hazard function follows the 45-degree baseline (figure
13). In conclusion, the final model was fitted the data successfully. In multivariate analysis, any
statistical test with a p< 0.05 was considered statistically significant. Then, associations were
summarized using an adjusted hazard ratio and statistical significances tested at 95%CI. Lastly,
the results were presented using texts, tables, and graphs.

4.12. Ethical Considerations

Ethical clearance was obtained from research and ethical review board of Nursing Department,
School of Nursing and Midwifery, College of Health Sciences, Addis Ababa University. A
formal letter of cooperation from the University was submitted to the hospital administrative
body for data collection (Dessie Referral Hospital). Permission was obtained from all concerned
bodies in the hospital. Next to these, searching and getting the chosen samples' medical record
was managed with the data collectors. Care was taken from disclosing patient’s records and
confidentiality was maintained by omitting their name and personal identification from the data
collection format. Finally, all collected informations was coded and locked in an isolated room
before entering the computer and locked by password after entering the computer.

4.13. Dissemination of the Result

The result will be submitted to Addis Ababa University, College of Health Sciences, School
of Nursing and Midwifery. Also, the result will be submitted to Dessie Referral Hospital.
Further effort will be made to present it at a workshop and conference and publish it on a
trustworthy journal.

20
 5. RESULTS
Among 590 neonatal cards, 542 records met the inclusion criteria in the final analysis with a
response rate of 91.86%.

5.1. Socio-Demographic Characteristics of the Study Participants

Of 542 study participants, 316 (58.3%) of them were males. More than half of the participants
(310 (57.2 %) came from rural areas. Most neonates (401 (73.99%) were admitted to the NICU
within 24 hrs of age. The cohort’s mean neonatal age at admission to NICU was 2.05 ± 0.12 SD
days (95%CI: 1.81, 2.27). The majority of mothers (451(83.21%) were belonging to 20-35
years. Among admitted neonates, 45 (8.3%) of them were in the GA group of very premature
(28-32 weeks), of which 33 (73.33%) died, which was more than 3 and 5 times higher than
premature (21.25%) and mature neonate (12.95%), respectively. The mean GA was
37.55±0.123 SD weeks (95%CI: 37.3, 37.9). The minimum birth weight was 800 gram and the
maximum was 4500 gram with a mean weight of 2651.11±32.56 SD gm (95%CI: 2587.15,
2715.06). Among 542 neonates, 181 were born with <2500 grams of birth weight, which
contributes more than half (61 (58.65%) of the total neonatal death.

Table 2: Socio-demographic characteristics of neonate and their mother in Dessie Referral

Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020.

Variable Category Total (%) Survival Status

Died (%) Censored (%)
Sex Male 316 (58.30) 60 (57.7) 256 (58.45)
Female 226 (41.70) 44 (42.3) 182 (41.55)
Residency Rural 310 (57.20) 63 (60.58) 247 (56.39)
Urban 232 (42.80) 41 (39.42) 191 (43.61)
Maternal age <20 years 56 (10.33) 12 (11.54) 44 (10.05)
20-35 years 451 (83.21) 84 (80.77) 367 (83.79)
>35 years 35 (6.46) 8 (7.69) 27 (6.16)
Neonatal age at <24 hrs. 401 (73.99) 82 (78.85) 319 (72.83)
admission 1-7 114 (21.03) 17 (16.34) 97 (22.15)
>7 day 27 ( 4.98) 5 (4.81) 22 (5.02)
GA in weeks 28-32 45 (8.30) 33 (31.73) 12 (2.74)

21
 33-36 80 (14.76) 17 (16.35) 62 (14.15)
≥37 417 (76.94) 54 (51.92) 363 (82.88)
Birth weight in <1000 7 (1.29) 6 (5.77) 1 (0.23)
gram 1000-1499 35 (6.46) 25 (24.04) 10 (2.28)
1500-2499 139 (25.65) 30 (28.84) 109 (24.89)
≥2500 361 (66.61) 43 (41.35) 318 (72.60)

5.2. Obstetric and Gynecological Characteristics of the Study Participants

In this study, 420 (77.49%) of neonates were born through SVD, which contributed 78 (75%)
to the total neonatal death. More than half 346 (63.84%) of the study participants had delivered
in Hospital. Most (528 (97.42%) of the deliveries were conducted by health professionals.
Regarding maternal obstetric and gynecological factors, 104 (19.19%) had PROM, 25 (4.61%)
had prolonged labor, and 31 (5.72%) had preeclampsia. More than half 20 (64.52%) of neonates
born from preeclampsia mothers died during the follow-up period. Among 104 dead neonates,
60 (57.69%) died from their mother’s parity of less than 2. Of the total neonates enrolled in this
study, 49 (9.04%) were twin pregnant. This study also showed that neonates born from mothers
who had no ANC follow-up were more vulnerable to death (33.33%) than their counterparts
(19.03%). (See Table 3).

Table 3: Obstetric and gynecological characteristics of the study participant in NICU of Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020.

Variable Category Total (%) Survival Status

Died (%) Censored (%)
Mode of SVD 420 (77.49) 78 (75) 342 (78.08)
delivery Instrumental 36 (6.64) 10 (9.62) 26 (5.94)
Caesarian section 86 (15.87) 16 (15.38) 70 (15.98)
Place of Hospital 346 (63.84) 62 (59.61) 284 (64.84)
delivery Health center 182 (33.58) 38 (36.54) 144 (32.88)
Home 14 (2.58) 4 (3.85) 10 (2.28)
Health professional 528 (97.42) 100 (96.15) 428 (97.72)

22
 Delivery Traditional birth 2 (0.37) 1 1
conducted by attendance
Other 12 (2.21) 3 (2.88) 9 (2.05)
PROM Yes 104 (19.19) 24 (23.08) 80 (18.26)
No 438 (80.81) 80 (76.92) 358 (81.74)
Prolonged Yes 25 (4.61) 12 (11.54) 13 (2.97)
labor No 517 (95.39) 92 (88.46) 425 (97.03)
Preeclampsia Yes 31 (5.72) 20 (19.23) 11 (2.51)
No 511 (94.28) 84 (80.77) 427 (97.49)
Placenta Previa Yes 3 (0.55) 2 1
No 539 (99.45) 102 (98.08) 437 (99.77)
Number of <2 217 (40.04) 60 (57.69) 157 (35.85)
parity 2-4 281 (51.85) 40 (38.46) 241 (55.02)
>4 44 (8.12) 4 (3.85) 40 (9.13)
Multiple Yes 49 (9.04) 13 (12.5) 36 (8.22)
pregnancies No 493 (90.96) 91 (87.5) 402 (91.78)
ANC follow-up Yes 536 (98.89) 102 (98.08) 434 (99.09)
No 6 (1.11) 2 (1.92) 4 (0.91)

5.3 Maternal Medical Diagnosis of the Study Participants

Among the total mothers enrolled in this study, 14 (2.85%) mothers had HIV/AIDS, and 10
(1.85%) had hepatitis B virus (HBV), but no death was recorded among those mothers neonate.
Besides, 10 (1.85%) of mothers had DM, in which 7 (70%) of their neonates died. Furthermore,
2 (0.37%) of mothers had epilepsy, of which all of their neonates died. (See table 4)

Table 4: Maternal medical diagnosis in NICU of Dessie Referral Hospital, Northeast Ethiopia,
January 1st, 2018 – December 31th, 2020.

Variable Category Total (%) Survival status

Died (%) Censored (%)
HIV/AIDS Yes 14 (2.58) 14 (3.17)
No 528 (97.42) 104(100) 424 (96.80)

23
 Hypertension Yes 4 (0.74) 4 (0.91)
No 538 (99.26) 104 (100) 434 (99.07)
Maternal Yes 2 (0.37) 1 (0.96) 1 (0.23)
anemia No 540 (99.63) 103 (99.04) 437 (99.77)
DM Yes 10 (1.85) 7 (6.73) 3 (0.68)
No 532 (98.15) 97 (93.27) 435 (99.32)
HBV Yes 10 (1.85) 10 (2.28)
No 532 (98.15) 104 (100) 428 (97.72)
CHF Yes 2 (0.37) 2 (0.46)
No 540 (99.63) 104 (100) 436 (99.54)
Epilepsy Yes 2 (0.37) 2 (1.92)
No 540 (99.63) 100 (98.08) 437 (100)

5.4 Common Medical and Other Diagnosis of Neonates Admitted to NICU

The significant admission medical problems documented among neonates admitted to NICU
during the follow-up period were low APGAR score (1st and 5th minute APGAR score <7, 242
(44.65%) and, 128 (20.66%) respectively), sepsis 143 (26.38%), PNA 68 (12.73%), Jaundice
36 (6.83%), MAS 42 (7.75%), and RDS 43 (7.93%). Among admitted neonates with RDS, 32
(74.42%) died, and RDS was the leading cause of death. In this study, sepsis was among a minor
killer medical problem, in which 10 (6.7%) of neonates died from 143 neonates. Other common
causes of admissions were Bag and mask resuscitation at birth, hypothermia, hypoglycemia,
anemia, congenital malformation. (See Table 5)

Table 5: Common medical and other diagnoses of neonates that were admitted to NICU of
Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020.

Variable Category Total (%) Survival Status

Died (%) Censored (%)
RDS Yes 43 (7.93) 32 (30.77) 11 (2.51)
No 499 (92.07) 72 (69.23) 427 (97.49)
Jaundice Yes 36 (6.83) 8 (7.69) 28 (6.39)
No 507 (93.17) 96 (92.31) 410 (93.61)

24
PNA Yes 68 (12.73) 23 (22.12) 45 (10.27)
No 473 (87.27) 81 (77.88) 393 (89.73)
MAS Yes 42 (7.75) 7 (6.73) 35 (7.99)
No 500 (92.25) 97 (93.27) 403 (92.01)
Sepsis Yes 143 (26.38) 10 (9.62) 133 (30.37)
No 399 (73.62) 94 (90.38) 305 (69.63)
Hypothermia Yes 70 (12.92) 6 (5.77) 64 (14.61)
No 472 (87.08) 98 (94.23) 374 (85.39)
Hypoglycemia Yes 46 (8.49) 1 (0.96) 45 (10.27)
No 496 (91.51) 103 (99.04) 393 (89.73)
Anemia Yes 18 (3.32) 8 (7.69) 10 (2.28)
No 524 (96.68) 96 (92.31) 428 (97.72)
Congenital Yes 33 (6.09) 4 (3.85) 29 (6.62)
malformation No 509 (93.91) 100 (96.15) 409 (93.38)
Birth trauma Yes 16 (2.95) 5 (4.81) 11 (2.51)
No 526 (97.05) 99 (95.19) 427 (97.49)
Bag and mask Yes 106 (19.56) 39 (37.5) 67 (15.30)
resuscitation No 436 (80.44) 65 (62.5) 371 (84.70)
Fist minute <7 242 (44.65) 71 (68.27) 171 (39.04)
APGAR score ≥7 300 (55.35) 33 (31.73) 267 (60.96)
Fifth minute <7 128 (20.66) 52 (50) 76 (17.35)
APGAR score ≥ 414 (79.34) 52 (50) 362 (82.65)
Initiation of BF Yes 226 (41.70) 18 (17.31) 208 (47.49)
within 1 hr No 316 (58.3) 86 (82.69) 230 (52.51)

25
5.5. Survival Status of Neonate
A total of 542 neonates that were admitted to NICU had been followed from 0 to 28 days. The
overall mean hospital stay was 5.631 (95%CI: 5.316, 5.946) with a minimum and maximum
follow-up time of 1 and 20 days. In this finding, 104 (19.19%) of the study participant died
during the follow-up period. And from this male accounts more than half of the death 60 (57.
69%). From neonates included in the analysis, 438 (80.81%) were censored (recovered,
discharged for left against medical advice, and referred to other health facilities) at the end of
the follow-up. (See figure 3 below).

Figure: 3 Outcome of neonates admitted to NICU of Dessie Referral Hospital, Northeast

Ethiopia, January 1st, 2018 – December 31th, 2020.
N.B*LAMA=Left against medical advice

26
5.5.1. Death/Failure of neonates and overall Kaplan- Meier failure function
The total follow-up was 2926 person-day in this study, with an incidence rate of 35.54 deaths
per 1000 person-day observations (95%CI: 29.33, 40.08). Neonates admitted in the first 24 hrs
were more likely to die with an incidence rate of 39.61/1000 person-time observations (95%CI:
30.87, 47.60). Incidence of neonatal death was higher in neonates born as very premature (28-
32 weeks of GA) with an incidence rate of 107.74/1000 neonate-days observations
(95%CI=76.19, 152.36), which is more than three and four times compared to neonates born as
premature (33-36 weeks of GA) and mature (>37 weeks) with an incidence rate of 34.29/1000
neonate-days (95%CI: 21.60, 54.42) and 24.22/1000 neonate-days observations (95%CI: 18.55,
31.62), respectively. Similarly, EVLBW and VLBW neonates had a high incidence of neonatal
death with 127.66/1000 neonate-days (95%CI: 57.35, 284.15) and 100.04/1000 neonate-days
(95%CI: 67.84, 148.59), respectively compared to LBW 37.83/1000 (95%CI: 26.45, 54.11) and
NBW 21.91/1000 (95%CI: 16.25, 29.54).

On the 1st day of hospital stay, the hazard probability of neonates was 11 (10.58%) with an
incidence rate of 20.29 death per 1000 neonate-days observations (95%CI: 11.24, 36.65).
Similarly, the failure probability of neonates at 3rd, 7th, and 14th days of hospital stay were 39
(37.5%), 83 (79.81%), and 103 (99.04%) with an incidence rate of 27.66 (95%CI: 20.21, 37.86),
32.79 (95%CI: 26.45, 40.66), and 35.56 (95%CI: 28.35, 42.21) deaths per 1000 person-days
observations, respectively. As shown in the figure below, the Kaplan Meier failure curve
increases stepwise during the hospital stay, and there is no death after 16 days of hospital stay.
(Figure 4).

27
Figure 4: Overall Kaplan-Meier failure estimate of neonate admitted to NICU of Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020.

5.5.2. Overall Kaplan- Meier Survival Function

The overall Kaplan- Meier estimate showed that the probability of survival of neonates was high
on the 1st day of admission, which comparatively decreases as follow-up time increases (see
figure 5). During the 1st day of hospital stay, a maximum (97.83%) (95%CI: 96.12, 98.80)
probability of survival was observed with a standard error of 0.65. The overall median survival
time of neonates admitted to NICU in the study was 14 days (95%CI: 12,) with a standard error
of 1.511. This study also showed that the probability of neonatal survival at the 7th and 14th day
of hospital stay was 75.75% (95%CI: 70.49, 80.20, SD = 2.47) and 42.12% (95%CI: 26.36,
57.08, SD = 8.05), respectively. At the 20 days of hospital stay, the overall survival probability
of neonates was 36.10% (95%CI: 19.51, 52.98) with a standard error of 8.87. (See figure 5).

28
Figure 5: Overall Kaplan-Meier survival estimate of neonate admitted to NICU of Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020.

5.5.3. Comparison of survival time for different groups of variables

The Kaplan-Meier estimator is a non-parametric statistic used to estimate the survival function
(the probability of survival) of right-censored data. Separate graphs Kaplan-Meier survivor
functions are used to compare the probability of survival among different categories of variables,
and it was constructed as described below. In general, the category defined by the upper curve
has a better survival probability than the category defined by the lower curve. The same for the
Kaplan-Meier failure (hazard) function is that the category defined by the upper curve has a
higher hazard probability than the category defined by the lower curve.

However, the statistical question is whether the observed difference seen on the graph is
significant or not. The Cochran-Mantel Haenszel Log-rank test can solve this. The Log-rank test
was performed to test the equality of survival curves of different categorical explanatory
variables (Table 6). The test statistics with a p-value <0.05 showed that there is a significant
difference in survival time for different categorical variables. These variables include neonatal

29
age at admission, GA, birth weight, number of parity, prolonged labor, preeclampsia, placenta
previa, maternal DM, maternal epilepsy, RDS, PNA, sepsis, neonatal anemia, Bag and Mask
resuscitation, 1st and 5th minute APGAR score, and time of initiation of breastfeeding.

As showed from the figure plotted below, the Kaplan Maier survival function for prolonged
labor had a lower survival time (median survival time of 7 days, 95%CI: 3,) than the counterparts
(median survival time of 14 days, (95%CI: 13,) with overall survival probability of 28.71% and
37.19% at the end of the study period respectively. The difference was statistically significant
at p-value = 0.0001 (See figure 6)

Figure 6: The Kaplan-Meier survival curves compare survival time of neonate with groups of
prolonged labor at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 –
December 31th, 2020

30
Likewise, neonates diagnosed with RDS had a lower survival time with 7 days (95% CI: 4, 8)
of median survival time compared to those who had not been diagnosed with RDS. At the 20
days of hospital stay, the overall survival of neonates with RDS and without RDS was 14.29%
and 39.37%, respectively (Figure 7). This survival time difference was statistically significant
with a p-value = 0.0000 (see Table 7).

Figure 7: The Kaplan-Meier survival curves compare survival time of neonate with groups of
RDS at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December
31th, 2020.

31
On the contrary, neonates diagnosed with sepsis had a longer survival time compared to those
who hadn’t sepsis, with overall survival status of 64.63% and 25.97% at the end of the follow-
up period. P-value = 0.0000 (See Figure 8).

Figure 8: The Kaplan-Meier survival curves compare survival time of neonate with groups of
sepsis at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December
31th, 2020.

32
Also, neonates who had Anemia had a lower survival time with a median survival time of 7 days
of age (95%CI: 3,) compared to the counterparts (median survival day 14, 95%CI: 13,) with
29.03% and 36.73% of survival probability at the end of the follow-up period. The difference
was statistically significant at p-value=0.0000 (see figure 9)

Figure 9: The Kaplan-Meier survival curves compare survival time of neonate with groups of
anemia at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 –
December 31th, 2020.

33
The median survival time of neonates who had a Bag and Mask resuscitation at birth was 12
days, but it was 14 days for neonates who had no Bag and Mask resuscitation at the time of
birth. The variation was significant at p-value =0.0001) (see fig 10).

Figure 10: The Kaplan-Meier survival curves compare survival time of neonate with groups of
bag and mask resuscitation at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January
1st, 2018 – December 31th, 2020.

34
The result of this study showed that neonates with a 1st minute APGAR scores of ≥7 had good
survival probability (median survival time of 16 days, 95%CI: 14,) than neonates with a 1st
minute APGAR scores of <7 (median survival time of 13 days, 95%CI: 12,). At the end of the
follow-up period, the overall survival probability of neonates with 1st minute APGAR score ≥7
and <7 was 38.00% and 33.06%. The difference was significant at P-value=0.0002. (See figure
11 below)

Figure 11: The Kaplan-Meier survival curves compare survival time of neonate with groups of
first minute APGAR score at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January
1st, 2018 – December 31th, 2020

35
Similarly, neonates with a 5th minute APGAR score of ≥ 7 had a longer survival time (median
survival time of 14 days, 95%CI: 14,) than neonates with a 5th minute Apgar score of <7 (median
survival time of 9 days, 95%CI: 7,). The difference was statistically significant at p-value =
0.0000 (see Figure 12)

Figure 12: The Kaplan-Meier failure curves compare survival time of neonate with groups of
fifth minute APGAR score at NICU, in Dessie Referral Hospital, Northeast Ethiopia, January
1st, 2018 – December 31th, 2020

36
Table 6: Median survival time and log-rank test for equality of survivor functions among
neonate admitted to NICU, in Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018 –
December 31th, 2020.

Variable Category Median Log-rank test P-value

survival time (x2)
(95%CI)
<20 years (8,) 0.28 0.8702
Maternal age 20-35 years 14 (12,)
>35 years 10 (8,)
Residence Rural 14 (12,) 0.79 0.3741
Urban (11,)
Sex Male 16 (12,) 0.02 0.9017
Female 14 (11,)
ANC follow up No (2,) 1.34 0.2464
Yes 14 (12,)
Multiple No 14 (12,) 1.53 0.2159
pregnancies Yes 16 (7,)
Neonatal age at <24 hrs 13 (12,) 6.24 0.045
admission 1-7 days 16 (14,)
>7 days (6,)
Mode of SVD 14 (12,) 1.22 0.5434
delivery Instrumental 9 (7, )
C/S (9,)

Place of Hospital 16 (14,) 2.23 0.3279

delivery Health center 12 (10,)
Home (4,)

GA in weeks <37 11 (8, 14) 20.15 0.0000

≥37 ----(14,)
Birth weight in <2500 11 (8, 16) 23.02 0.0000
gram ≥2500 ----(14,)

37
DM Yes 6 (3,) 15.53 0.0001
No 14 (13,)
Epilepsy Yes 1 (1,) 11.51 0.0007
No 14 (12,)
RDS Yes 7(4,8) 51.18 0.0000
No 16 (14,)
PNA Yes 12 (9,) 4.76 0.029
No 14 (13,)
Sepsis Yes ----(14,) 28.61 0.0000
No 12 (10,16)
Anemia Yes 7 (3,) 17.59 0.0000
No 14 (13,)
Bag and Mask Yes 12 (8,) 14.81 0.0001
resuscitation No 14 (14,)
First minute <7 13 (12,) 13.73 0.0002
APGAR score ≥7 16 (14,)
Fifth minute <7 9 (7,) 29.60 0.0000
APGAR score ≥7 14 (14,)
Initiation of BF Yes 13.42 0.0002
within 1 hr No 13 (12, 16)
Number of <2 11 (8,14) 15.82 0.0001
parity ≥2 16 (14,)
Prolonged Yes 7 (3,) 20.95 0.0001
labor No 14 (13,)
PROM Yes (10, ) 0.35 0.5528
No 14 (12,)

Placenta previa Yes 5 (5,) 5.84 0.0156

No 14 (12,)
Preeclampsia Yes 7 (5,8) 29.01 0.0000
No 16 (13,)

38
5.6. Predictors of Time to Death among Neonates Admitted to NICU
The relationship between the independent and outcome variables was analyzed using Cox
proportional hazard regression model. In the bivariate Cox proportional regression analysis, the
number of parity, prolonged labor, preeclampsia, placenta previa, epilepsy, GA, birth weight,
RDS, PNA, neonatal anemia, bag and mask resuscitation, 1st and 5th minute APGAR score,
and time of initiation of breastfeeding were statistically significant predictors of time to death
of neonates (p-value < 0.05). Furthermore, to find independent predictors of time to death of
neonates and survival, multivariate cox regression analysis was done for all predictors with a p-
value <0.25 in the bivariate analysis. Nevertheless, the strong predictors of time to death in the
multivariate analysis were found that the number of parity, prolonged labor, preeclampsia, DM,
birth weight, PNA, and neonatal anemia.

Neonates born from primipara mothers had approximately 2 times more likely to die (AHR
2.231 (95%CI: 1.442, 3.451) than neonates born from multiparous mothers. The hazard ratio for
neonates born with prolonged labor was more than 3 times more likely than those neonates born
without prolonged labor (AHR: 3.317 (95%CI: 1.462, 7.526). The hazard ratio for death was
2.165 times higher among neonates born from mothers who have preeclampsia during their
pregnancy as compared to their counterparts (AHR: 2.165 (95%CI: 1.102, 4.253). Neonates
born from diabetic mothers were approximately 4 times more likely to die than those born from
mothers without DM (AHR: 3.738 (95%CI: 1.569, 8.904). The hazard of neonatal mortality was
3.026 times higher for those neonates born with low birth weight as compared to neonates born
with normal birth weight (AHR: 3.026 (95%CI: 1.526, 5.999). Similarly, neonates admitted to
NICU with PNA at baseline had 2.613 times the risk of death than those admitted to NICU
without PNA (AHR: 2.613 (95%CI: 1.114, 6.126). The hazard of death was 9.139 times higher
among neonates admitted with the diagnosis of neonatal anemia than those neonates admitted
without anemia (AHR: 9.139 (95%CI: 3.329, 25.084), at P-value=0.000.

39
Table 7: Results of bivariate and multivariate analysis using Cox regression model for predictors
of time to death among neonate admitted to NICU, in Dessie Referral Hospital, Northeast
Ethiopia, January 1 st, 2018 – December 31th, 2020.

Variable Category CHR (95%CI) P- AHR (95%CI) P-

value value
Number of <2 2.042 (1.383, 3.017) 0.000 2.231 (1.442,3.451) 0.000
parity ≥2 1
ANC follow Yes 1
up No 2.437(0.599, 9.915) 0.213 1.234 (0.262, 5.812) 0.790
Prolonged Yes 3.070(1.661, 5.553) 0.000 3.317 (1.462,7.526) 0.004
labor No 1
Preeclampsia Yes 3.747(2.292,6.127) 0.000 2.165 (1.102,4.253) 0.025
No 1
Placenta Yes 5.021(1.229,20.512) 0.025 0.562 (0.100, 3.160) 0.513
previa No 1
Place of Hospital 1
delivery Health 1.316 (0.879, 1.972) 0.182 1.082 (0.682, 1.717) 0.738
center
Home 2.444(0.888, 6.731) 0.084 1.488 (0.469, 4.728) 0.500
Maternal Yes 4.106(0.570,29.564) 0.161 2.855(0.356,22.881) 0.324
anemia No 1
DM Yes 3.363(1.555, 7.270) 0.002 3.738 (1.569,8.904) 0.003
No 1
Epilepsy Yes 5.724(2.099,15.609) 0.001 4.422(0.791,24.726) 0.091
No 1
GA at birth <37 2.457 (1.671, 3.614) 0.000 1.112 (0.462, 2.677) 0.813
in weeks ≥37 1
Birth weight <2500 2.553 (1.727, 3.775) 0.000 3.026 (1.526, 5.999) 0.002
in gram ≥2500 1
RDS Yes 4.010 (2.638,6.097) 0.000 2.017 (0.931,3.759) 0.075

40
 No 1
PNA Yes 1.656 (1.041,2.635) 0.033 2.613 (1.114, 6.126) 0.027
No 1
Sepsis Yes 0.204 (0.106,0.392) 0.000 0.659 (0.297, 1.462) 0.305
No 1
Neonatal Yes 4.467 (2.142, 9.316) 0.000 9.139(3.329,25.084) 0.000
anemia No 1
Birth trauma Yes 1.823 (0.741, 4.482) 0.191 1.889 (0.551,6.483) 0.312
No 1
Bag and Yes 2.132 (1.432,3.173) 0.000 1.510 (0.758, 3.003) 0.241
mask No 1
resuscitation
1st minute <7 2.209 (1.460,3.341) 0.000 0.967 (0.484, 1.931) 0.924
APGAR ≥7 1
score
5th minute <7 2.789 (1.898,4.099) 0.000 1.243 (0.605, 2.555) 0.554
APGAR ≥7 1
score
Initiation of Yes 1
BF within 1 No 2.538 (1.524, 4.225) 0.000 0.852 (0.394, 1.844) 0.685
hr

Those bold numbers above showed significantly associated predictors of time to death.
CHR: Crude Hazard Ratio; AHR: Adjusted Hazard Ratio; CI: Confidence Interval
P-value: level of significance at <0.05

41
5.7. Multicollinearity Test
Predictor variable in a multiple Cox-regression Model can linearly predict from the others with
a considerable degree of accuracy. Multicollinearity occurs when predictor covariates in the
Cox-regression model are correlated. This correlation is a problem because the relationship
between each predictor variable with the outcome variable should be independent.
Multicollinearity can be checked by calculating variance inflation factor (VIF) for each
predictor variable (based on binary cox-regression outcome). According to some guidelines, a
VIF value less than 10 is considered as an acceptable collinearity level. However, a VIF value
greater than 4 needs further investigation. In this study the maximum VIF value was 3.08 with
a mean VIF of 1.62. Therefore, there was no multicollinearity, and we can interpret the results
of this study. (See table 9).

Table 9: Multicollinearity check for each variable based on binary outcome (p-value<0.25) to
neonates admitted in NICU at Dessie Referral Hospital, Northeast Ethiopia, January 1st, 2018
– December 31th, 2020

Variable VIF 1/VIF

5th minute APGAR score 3.08 0.324580
Bag and Mask resuscitation 3.06 0.326814
GA at birth in weeks 2.80 0.357054
PNA 2.16 0.462119
Birth weight 2.09 0.477332
RDS 2.02 0.494866
1st minute APGAR score 1.97 0.508196
Initiation of BF within 1 hr 1.69 0.591746
Preeclampsia 1.44 0.694875
Sepsis 1.23 0.810378
Neonatal anemia 1.20 0.833318
Prolonged labor 1.18 0.845737
Birth trauma 1.15 0.871025
Place of delivery 1.10 0.908991
Number of parity 1.07 0.936072

42
 Epilepsy 1.06 0.942373
ANC follow up 1.06 0.946078
Placenta previa 1.03 0.968156
Maternal anemia 1.03 0.974988
DM 1.02 0.980894
Mean VIF 1.62

5.8. Test of Proportional Hazard Assumption

Based on binary and multicollinearity outcomes, testing proportional hazard assumption is vital
to interpret, use fitted proportional hazard models, and accept multivariate analysis results. In
this study, Schoenfeld residuals proportional hazard assumption test for the individual
covariates was done and global tests were used. If P-value < 0.05, then the proportional hazard
assumption is rejected. (See table 7). Also, a concordance probability test was performed
(Harrell’s c=0.8286).

Table 8: Schoenfeld Residuals test for Cox-Proportional Hazard Regression model assumption
(estat, phtest) of each variable for neonates admitted in NICU at Dessie Referral Hospital,
Northeast Ethiopia, January 1st, 2018 – December 31th, 2020.

VARIABLES RHO CHI2 (X2) P-VALUE

Parity -0.02547 0.07 0.7926
ANC follow up -0.01834 0.03 0.8699
Prolonged labor 0.01415 0.02 0.8784
Preeclampsia 0.10112 1.21 0.2717
Placenta previa 0.02558 0.09 0.7656
Within 1 hr BF initiation 0.13377 1.42 0.2333
Place of delivery (HC) 0.04399 0.20 0.6539
Place of delivery (Home) 0.08676 0.56 0.4534
Maternal anemia 0.12809 1.70 0.1928
DM 0.06052 0.44 0.5059
Epilepsy -0.01076 0.01 0.9164
GA 0.00997 0.01 0.9230
Birth weight 0.00162 0.00 0.9867

43
RDS -0.08130 0.78 0.3756
PNA -0.04986 0.34 0.5572
Sepsis 0.02364 0.05 0.8230
Neonatal anemia -0.12699 1.46 0.2270
Birth trauma -0.02825 0.06 0.8064
Bag and mask -0.10275 1.13 0.2869
resuscitation
1st minute Apgar score -0.14225 2.07 0.1501
5th minute Apgar score 0.01955 0.04 0.8363
GLOBAL TEST 20.89 0.4655

44
5.9. Cox-Snell residuals Test
Cox-Snell residuals test is used to assess the overall goodness of fit in survival models. It is done
graphically with the usual graphs of Cox-Snell, and it is observed that residuals from a correctly
fitted model follow exponential unite distribution along the 45 degree of slope (baseline). It is
used to identify extreme observations that need additional investigation. In this study, the hazard
function follows the 45-degrees line closed to the baseline. It showed that the multivariate cox-
regression model is fitted for the analysis and interpretation of the results of this study.

Fig 13: Cox-Snell residual cumulative hazard graph on neonates admitted in NICU at Dessie
Referral Hospital, Northeast Ethiopia, January 1st, 2018 – December 31th, 2020

45
 6. DISCUSSION
This retrospective follow-up study was aimed to determine the survival status and predictors of
mortality among neonates admitted to NICU in Dessie Referral Hospital. This study shows that
the overall neonatal mortality (NMR) in NICU of Dessie Referral Hospital in the study period
was 104 (19.19%), which is equivalent to the NMR of 192/1000 admitted neonates (95%CI:
160, 228). This finding was more than three times higher than the study done in Ethiopia, Tigray
region (5.90% (62.5/1000 live birth) (63) and slightly in line with a study conducted in Arba
Minch General Hospital (20.8%) (31) and Debre Markos Referral Hospital (21.3%) (34). Also,
the finding was higher than a study done in northern Tanzania which was 10.7% (28) and
Nigeria (14.2%) (27), but lower than the study conducted in Mauritania (34.7%) (29). Similarly,
the neonatal mortality rate of this study was higher than studies conducted in Iran (11.4%) (22)
and India (10.4%) (24). These different results might be due to different reasons. The possible
reason may be different in methodology and study period. The other justification could be the
difference in availability of medical equipment and other resources in NICU, and the difference
in availability of trained human resources (difference in care).

At the end of the follow-up period, the overall incidence of neonatal mortality was 35.54 deaths
per 1000 neonate-day observations. This result was higher than the global and sub-Saharan
African incidence of neonatal mortality rate, 18 and 27 deaths per 1000 live births, respectively
(1). Also, this finding was higher than the 2019 report of mini EDHS (29 per 1000 live births)
(7). Likewise, the result of this study was higher than studies conducted in Ethiopia, Arba Minch
General Hospital (31.6 per 1,000 neonate days) (31), and Jimma University Medical center (30
deaths per 1,000 institutional live births) (10). This finding was in line with reports in
Afghanistan, which was 36 deaths/1000 live birth (25); however, it was lower than the
Democratic Republic of Congo (94 deaths per 1000 live births) (30). This apparent difference
could be related to several reasons, such as economic wealth difference. Developed countries
might have skilled health professionals with organized personnel and equipment in the delivery
room and NICU to give essential newborn care and treatment for those sick neonates. The other
reason might be geographical location and climate difference, in which hot and humid day
climate increases the risk of neonatal illness and mortality. Community educational level and
cultural difference might also be the other reasons.

46
In this study, the hazard probability of neonates in the 1st, 7th, and 14th days of hospital stay was
10.58%, 73.08%, and 99.04%, with an incidence rate of 20.29, 32.79, and 35.56 deaths per 1000
neonate-days observation, respectively. This finding was similar to a study done in UoGCSH in
which neonatal failure probability at the 1st, 7th, and 14th day of hospital stay was 13.79%,
83.91%, and 97.70% with an incidence rate of 23.81, 31.48, and 28.90 deaths/1000 neonate-
day observations, respectively (33). Generally, the highest failure probabilities occurred among
neonates admitted in the early neonatal period (99 (95.19%) with an incidence rate of
35.59/1000 neonate-days observations. The possible scientific explanation could be the neonatal
immunity system. Both their innate and adaptive immunity is functionally immature, resulting
in increased neonatal morbidity and mortality in the early neonatal period. The other possible
justification might be pregnancy and intrapartum-related complications, delaying early
identification and treatment of those complications. But, it was higher than a study in the
UoGCSH (84%) (33). The reason may be that the difference in the duration of follow-up time
(year) and study area.

The overall median survival time of neonates admitted to NICU was 14 days. It was higher than
a study in Arba Minch General Hospital (6 days) (31). The possible explanation for this
difference could be the sample size (N=332) and the study area. The other reason could be
changing (updating) of treatment modalities over time.

In the present study, the overall survival probability was 36.10% in the 20 days of the follow-
up period. The survival probability at the 1st, 7th, and 14th days of hospital stay was 97.83%,
75.75%, and 42.12%. This finding was consistent in the study done at Arba Minch General
Hospital in which the 1st, 7th, and 14th day of survival probabilities were 96.1%, 75%, and 69.9%,
respectively, with an overall survival probability of 66.2% (31).

After adjusting the multivariate cox proportional regression model, the significant independent
predictors of neonatal mortality in the NICU were low birth weight, primiparous mother,
prolonged labor, preeclampsia, maternal DM, PNA, and neonatal anemia.

Birth weight was an important independent predictor of neonatal survival in which neonates
with low birth weight had 3 times the hazard of death compared to neonates with normal birth
weight (AHR: 3.026, P=0.002). This finding is comparable with other studies conducted in
different places (9,24,29,37,38,40,41). The possible scientific justification may be that LBW

47
neonates are at potential risks of developing complications than normal birth weight. Secondly,
in this study, most LBW neonates were also premature with many physiologic problems,
especially with an immature lung leading to RDS secondary to HMD.

Being a primiparous mother was another significant predictor of the survival status of neonates.
The number of parity<2 was increased the hazard of death by more than two folds than the
number of parity≥2 (AHR=2.231, P=0.000). This result was in agreement with other studies
conducted in Assosa Zone and Ayder Comprehensive Hospital (50,51). It may be because
primipara mothers are more likely to experience complications during pregnancy and birth.

Having prolonged labor was also found to increase the hazard of neonatal death (AHR: 3.317,
P=0.004), which was similar to a study conducted in Arba Minch General hospital (31). The
possible scientific justification might be because prolonged labor is the primary cause of
neonatal infection, perinatal asphyxia, and hypoxemia, contributing to the development of
neonatal morbidity and mortality.

The hazard of death for neonates born from mothers with preeclampsia was 2.165 times higher
compared to those born from mothers without preeclampsia during their pregnancy (AHR:
2.165, P=0.025). This is due to preeclampsia affecting the mother’s blood circulation, thereby
affecting blood flow to the placenta, leading to fetal hypoxemia in the uterus. Secondly,
preeclampsia can be threatening to maternal life and in this case, terminating the pregnancy is
mandatory, leading to the delivery of premature babies with immature organ systems to survive.

Neonates born from diabetic mothers had close to 4 times a risk of death than neonates born
from mothers without DM (AHR: 3.738, P=0.003). This might be due to complications of DM
for mothers and their neonates as stated by many studies and literature (60,61). The possible
scientific justification might be that neonates born from diabetic mothers are often larger than
other neonates, making vaginal delivery difficult and may increase the risk of nerve injury and
other trauma during delivery that increases the risk of neonatal death. The other scientific
justification might be that a pregnant diabetic mother may have preeclampsia, leading to the
delivery of a premature baby, which was the leading cause of neonatal mortality in this study.

Having PNA at admission was another vital predictor of time to neonatal death. Neonates who
were diagnosed with PNA at admission had approximately 3 times a hazard of death than

48
neonates who hadn’t PNA at admission (AHR: 2.613, P= 0.027). Similar results have been
recorded in Ethiopia and other studies (24,40,48,56,57). The possible scientific justification
could be, asphyxia contributes to the deprivation of oxygen in the body, leading to hypoxia,
hypercarbia, and acidosis formation, which eventually damages vital organs leading to death.

The last significant variable that predicted the time to death of neonates in this study was
neonatal anemia. Neonates who had anemia during admission were almost nine times more
likely to die than neonates who hadn’t anemia (AHR: 9.139, P=0.000). This might be because
of the inability of RBC to deliver adequate oxygen to the tissue of the neonate.

49
 7. LIMITATION AND STRENGTH OF THE STUDY
Strength

The data were collected by nurses who trained for neonatal care in NICU, which has an essential
role in ensuring data quality.

The data were gathered from three consecutive years of admission/observations with
proportional allocation, which might increase the number of events and decrease variability.

Considering censored observations was also an important strength of this study since it provides
a more accurate estimation for survival analysis.

It gives an insight for researchers, especially for a prospective study.

It was easy to make a sequential association between outcome covariate (death) and predictor
covariates, which were recorded at admission.

Limitations

Since the study was conducted using secondary sources, some critical covariates obtain from
the mother were missing. These might be significant predictors of neonatal mortality such
educational status, occupational status, nutritional status, marital status, economic status, birth
interval, history of abortion, and history of stillbirth.

The study area covers only Dessie Referral Hospital. As a result, its generalizability to all
hospitals of Ethiopia may be impossible.

Medical records that were incomplete or lost at the time of data collection were excluded. These
possibly introduce selection bias that results under or overestimation of the findings of the study.

Unavailability of reference study for median survival time of neonates admitted in NICU and
for some variables that were independent predictors of death in this study.

50
 8. CONCLUSION
In the present study, 104 (19.19%) neonates died during the follow-up period. The overall
incidence of NMR was 35.54 deaths per 1000-neonate day observation with an overall median
survival time of 14 days. The incidence of death was high in the early neonatal period. Also, the
overall cumulative survival probability was 36.10% at the end of the study period. This study
also showed that the independent predictors of time to death of neonates admitted to NICU were
low birth weight, primiparous mother, prolonged labor, preeclampsia, maternal DM, PNA, and
neonatal anemia.

51
 9. RECOMMENDATION
Even though quality of health care service expanded, especially for maternal and newborn care,
neonatal mortality is still a concerning issue in all Ethiopia’s study areas. Based on the findings
of this study, I forwarded the following recommendations.

To the federal minister of health

Researches, including this study, showed that incidence of neonatal mortality had no reduction
progress, especially since 2016 till now. Therefore, federal health policymakers should revise,
particularly maternal and neonatal health policies, based on up-to-date study findings to address
interventions for predictors of neonatal mortality

To Dessie Referral Hospital

The Hospital should give an early diagnosis, early intervention (care), follow-up, and regular
monitoring for mothers with preeclampsia during their pregnancy since preeclampsia was an
independent predictor of neonatal death in the current study.

Another emphasis should be given to prolonged labor, and delivery should be conducted within
18 hrs of the onset of labor to reduce the incidence of neonatal mortality.

The hospital should facilitate more research to find out more detailed diagnoses of predictors of
neonatal mortality.

To health care providers of Dessie Referral Hospital

A special priority, monitoring and strict follow-up must be given to neonates admitted in their
first 7 days of life because this study found that the highest incidence of neonatal death was in
this time of neonatal age.

The health care provider should give especial attention and care for neonates born with
prolonged labor, primiparous mother, having preeclampsia mother, having DM mother, low
birth weight, and neonatal anemia.

It would be better to strengthen screening of DM during pregnancy

52
To future researchers

Upcoming researchers shall conduct a longitudinal prospective cohort study to incorporate

impossible variables in retrospective study like maternal socio-economic, genetic and
environmental factors.

The subsequent researchers shall also conduct a community-based study on the survival status
of neonates to avoid the estimation effect due to censored data.

53
 REFERENCE
1. UN IGME. Levels & Trends in Child Mortality: Estimates: Report 2018.
Who/Unicef/World Bank/Un. 2018;1–48.

2. Lawn JE, Kinney M, Blencowe H, The Lancet Every Newborn Study Group, Coghlan
N, Farber J. The Lancet’s Every Newborn Series. Lancet. 2014;1–8.

3. Ibor EK, Joan I, Offiong I, Chigozie U. Predictors of Neonatal Mortality in a Tertiary

Institution of a Developing Country. 2020;11(7).

4. Afolabi BM. Sub-Sahara African Neonates – Ghosts to Statistics. J Neonatal Biol.

2017;06(01):1–3.

5. Grady SC, Frake AN, Zhang Q, Bene M, Jordan DR, Vertalka J, et al. Neonatal
mortality in East Africa and West Africa: A geographic analysis of district-level
demographic and health survey data. Geospat Health. 2017;12(1).

6. Hug L, Alexander M, You D, Alkema L. National, regional, and global levels and
trends in neonatal mortality between 1990 and 2017, with scenario-based projections to
2030: a systematic analysis. Lancet Glob Heal [Internet]. 2019;7(6):e710–20. Available
from: http://dx.doi.org/10.1016/S2214-109X(19)30163-9

7. FEDERAL DEMOCRATIC REPUBLIC OF ETHIOPIA. Ethiopia Mini Demographic

and Health Survey (2019 EMDHS). In.

8. Mengistu BA, Yismaw AE, Azene ZN, Mihret MS. Incidence and predictors of
neonatal mortality among neonates admitted in Amhara regional state referral hospitals,
Ethiopia: Prospective follow up study. BMC Pediatr. 2020;20(1):1–14.

9. Desalew A, Sintayehu Y, Teferi N, Amare F, Geda B, Worku T. Cause and predictors

of neonatal mortality among neonates admitted to neonatal intensive care units of public
hospitals in eastern Ethiopia : a facility-based prospective follow-up study. 2020;1–11.

10. Seid SS, Ibro SA, Ahmed AA, Olani Akuma A, Reta EY, Haso TK, et al. <p>Causes
and factors associated with neonatal mortality in Neonatal Intensive Care Unit (NICU)
of Jimma University Medical Center, Jimma, South West Ethiopia</p>. Pediatr Heal
Med Ther. 2019;Volume 10:39–48.
54
11. Progress C, Forward M. REACHING THE EVERY NEWBORN NATIONAL 2020
MILESTONES COUNTRY PROGRESS , PLANS AND MOVING FORWARD. 2020.

12. UNICE. Neonatal mortality- The neonatal period is the most vulnerable time for a
child. 2019 Retrieved from https://data.unicef.org/topic/child-survival/neonatal-
mortality/. 2019;2019.

13. WHO. Newborns: improving survival and well-being, 19 September 2020 Retrieved
from https://www.who.int/news-room/fact-sheets/detail/newborns-reducing-mortality#.

14. FEDERAL DEMOCRATIC REPUBLIC OF ETHIOPIA. Demographic and Health

Survey 2016. In.

15. Bocquenet G, Chaiban T, Cook S, Escudero P, Franco A, Romo CG, et al. A fair
chance for every child. 2016.

16. WHO and UNICEF. EVERY NEWBORN: An Action Plan To End Preventable Deaths,
JUNE 2014.

17. Mehra V, Misunas C, Mizunoya S, Murphy P, Murray C, Petrowski N, et al.

PROGRESS FOR EVERY CHILD.

18. Ethiopian Public Health Institute (EPHI). COUNTDOWN TO 2015 : ETHIOPIA ’ S

PROGRESS TOWARDS REDUCTION IN UNDER-FIVE MORTALITY, 201 4
COUNTRY CASE STUDY.

19. Federal Ministry of Health Maternal and Child Health Directorate. National Strategy for
Newborn and Child Survival in Ethiopia (2015/16 – 2019/20). Addis Ababa, JUNE
2015. 2019;(June 2015).

20. The Federal Democratic Republic of Ethiopia Ministry of Health. Health Sector
Transformation Plan, 2015/16 - 2019/20 , October 2015.

21. Versantvoort JMD, Kleinhout MY, Ockhuijsen HDL, Bloemenkamp K, Vries WB De,
Hoogen A Van Den. Helping Babies Breathe and its effects on related stillbirths and
neonatal mortality in low- - resource settings : a systematic review. 2020;127–33.

22. Karimi P, Mahmudi L, Azami M, Badfar G. Mortality in Neonatal Intensive Care Units

55
 in Iran : A Systematic Review and Meta-Analysis. 2019;10(3).

23. Profile N, Morbidity N. Journal of Neonatal and Pediatric Morbidity and Mortality
Among Neonates Admitted to a Neonatal Intensive Care Unit of a Tertiary Care
Teaching Hospital of Jammu and Kashmir ( India ). 2018;(March).

24. Ravikumar SA, Elangovan H, Elayaraja K, K ASK, Sa R, Contemp IJ, et al. Morbidity
and mortality profile of neonates in a tertiary care centre in Tamil Nadu : a study from
South India. 2018;5(2):377–82.

25. Unicef. for every child. Maternal and Newborn Health Disparities in Afghanistan, 2015.

26. Hedstrom A, Ryman T, Otai C, Nyonyintono J, Mcadams RM, Lester D. Demographics

, clinical characteristics and neonatal outcomes in a rural Ugandan NICU. 2015;1–9.

27. Ekwochi U, Ndu IK, Nwokoye IC, Ezenwosu OU, Amadi OF, Osuorah DIC. Pattern of
morbidity and mortality of newborns admitted into the sick and special care baby unit
of Enugu State University Teaching Hospital , Enugu state. 2014;17(3):346–51.

28. Mmbaga BT, Lie RT, Olomi R, Mahande MJ, Kvåle G, Daltveit AK. Cause-specific
neonatal mortality in a neonatal care unit in Northern Tanzania : a registry based cohort
study. 2012;

29. Weddih A, Ahmed MLCB, Sidatt M, Abdelghader N, Abdelghader F, Ahmed A, et al.

Prevalence and factors associated with neonatal mortality among neonates hospitalized
at the national hospital Nouakchott, Mauritania. Pan Afr Med J. 2019;34.

30. Unicef. for every child. Maternal and Newborn Health Disparities in Democratic
Republic of the Congo, 2015.

31. Dessu S, Gebremeskel F, Alemu G, Seman B. Survival Status and Predictors of

Neonatal Mortality among Neonates Who were Admitted in Neonatal Intensive Care
Unit at Arba Minch General Hospital, Southern Ethiopia. Pediatr Ther. 2018;08(03).

32. Orsido TT, Asseffa NA, Berheto TM. Predictors of Neonatal mortality in Neonatal
intensive care unit at referral Hospital in Southern Ethiopia: A retrospective cohort
study. BMC Pregnancy Childbirth. 2019;19(1):1–9.

56
33. Gudayu TW, Zeleke EG, Lakew AM. <p>Time to Death and its Predictors among
Neonates Admitted in the Intensive Care Unit of the University of Gondar
Comprehensive Specialized Hospital, Northwest Ethiopia</p>. Res Reports Neonatol.
2020;Volume 10:1–10.

34. Alebel A, Wagnew F, Petrucka P, Tesema C, Moges NA, Ketema DB, et al. Neonatal
mortality in the neonatal intensive care unit of Debre Markos referral hospital ,
Northwest Ethiopia : a prospective cohort study. 2020;1–11.

35. UNITED NATIONS. Transforming our world: the 2030 agenda for sustainable
development A/RES/70/1.

36. Mengesha HG, Sahle BW. Cause of neonatal deaths in Northern Ethiopia : a
prospective cohort study. BMC Public Health [Internet]. 2017;1–8. Available from:
http://dx.doi.org/10.1186/s12889-016-3979-8

37. Onwuanaku CA, Okolo SN, Ige KO, Okpe SE, Toma BO. The effects of birth weight
and gender on neonatal mortality in north central Nigeria. 2011;2–6.

38. Sabzehei MK, Basiri B, Shokouhi M, Eghbalian F. Causes and Risk Factors Associated
to Neonatal Mortality in Neonatal Intensive Care Unit ( NICU ) in Besat Hospital
Hamadan-Iran in 2015 to 2016. 2018;6(57):8185–94.

39. Haghighi L, Nojomi M, Mohabbatian B, Najmi Z. Survival predictors of preterm

neonates: Hospital based study in Iran (2010-2011). Iran J Reprod Med.
2013;11(12):957–64.

40. Shahidullah M, Hasan MZ, Jahan I, Ahmed F, Chandra A. Perinatal Characteristics and
Outcome of Neonates at NICU of a Tertiary Level Hospital in Bangladesh.
2017;41(1):34–9.

41. Kahsay AH, Abebe HT, Gebretsadik LG, Tekle TH. Survival and predictors of early
neonatal death in neonatal intensive care unit of Mekelle general and Ayder
comprehensive specialized hospitals, Northern Ethiopia, 2018: prospective cohort
study. 2019;1–30.

42. Ballot DE, Chirwa T. Factors Associated with Survival to Discharge of Newborns in a

57
 Middle-Income Country. Epidemiol Commun Non-Communicable Dis - Attrib
Lifestyle Nat Humankind. 2016;

43. Hoque M, Haaq S, Islam R. Causes of neonatal admissions and deaths at a rural hospital
in KwaZulu-Natal , South Africa Causes of neonatal admissions and deaths at a rural
hospital in KwaZulu-Natal , South Africa. 2015;8782.

44. Eshete A. When Do Newborns Die ? Timing and Cause-Specific Neonatal Death in
Neonatal Intensive Care Unit at Referral Hospital in Gedeo Zone : A Prospective
Cohort Study. 2020;2020.

45. Rm M, Ngbonda D, Oj A. Predictive factors of neonatal mortality in intensive neonatal

care unit at Goma Eastern Democratic Republic of Congo. 2019;9(2):58–61.

46. Kozuki N, Lee ACC, Silveira MF, Sania A, Vogel JP, Adair L, et al. The associations
of parity and maternal age with small-for-gestational-age , preterm , and neonatal and
infant mortality : a meta-analysis. 2013;13(Suppl 3).

47. Dessu S, Kote M, Gebremeskel F, Girum T. Original article Predictors of Neonatal

mortality among Neonates who admitted in Neonatal intensive Care Unit at Arba Minch
General Hospital.

48. Bitew ZW, Alemu A, Ayele EG, Jember DA, Haile MT, Worku T. Incidence Density
Rate of Neonatal Mortality and Predictors in Sub-Saharan Africa : A Systematic
Review and Meta-Analysis. 2020;2020.

49. Li C, Yan H, Zeng L, Dibley MJ, Wang D. Predictors for neonatal death in the rural
areas of Shaanxi Province of Northwestern China : a cross-sectional study. 2015;1–8.

50. Kidus F, Woldemichael K, Hiko D. Predictors of neonatal mortality in Assosa zone ,

Western Ethiopia : a matched case control study. 2019;5:1–13.

51. Berhe AH. Prevalence and Factors Associated with Neonatal Mortality at Ayder
Comprehensive Specialized Hospital , Northern Ethiopia . A Cross-Sectional. 2020;

52. abay woday, Yohannes Mekuria CSD. Neonatal Mortality And Associated Factors
Among Neonates Admitted In Public Hospitals , Afar Pastoral Region : A Health

58
 Facility Based Study. :1–22.

53. Health P. Patterns of admission and factors associated with neonatal mortality among
neonates admitted to the neonatal intensive care unit of University of Gondar Hospital ,
Northwest Ethiopia. 2017;57–64.

54. Muhammed G, Kibria A, Burrowes V, Choudhury A, Sharmeen A, Ghosh S, et al.

Determinants of early neonatal mortality in Afghanistan : an analysis of the
Demographic and Health Survey 2015. 2018;1–12.

55. Miles M, Thi K, Dung K, Ha LT, Liem NT, Ha K, et al. The cause-specific morbidity
and mortality , and referral patterns of all neonates admitted to a tertiary referral
hospital in the northern provinces of Vietnam over a one year period. 2017;1–12.

56. atnafu Mekonnen Tekleab1 gesit Metaferia amaru1 Yemisrach abeje Tefera2. Reasons
for admission and neonatal outcome in the neonatal care unit of a tertiary care hospital
in Addis Ababa : a prospective study. 2016;17–23.

57. Alemu AY, Belay GM, Berhanu M, Minuye B. Determinants of neonatal mortality at
neonatal intensive care unit in Northeast Ethiopia : unmatched case-control study.
2020;9.

58. Bulo H. Effects of HIV infection on maternal and neonatal health in southern
Mozambique : A prospective cohort study after a decade of antiretroviral drugs roll out.
2017;1–17.

59. Hassan M, Begum M, Haque SMZ, Jahan N, Yasmeen BHN, Mannan A, et al.
Immediate Outcome of Neonates with Maternal Hypertensive Disorder of Pregnancy at
a Neonatal Intensive Care Unit. 2015;6(2):57–60.

60. Anjum SK, Yashodha HT. A study of neonatal outcome in infants born to diabetic
mothers at a tertiary care hospital. 2018;5(2):489–92.

61. Opara PI, Jaja T, Onubogu UC. OF PEDIATRICS Morbidity and mortality amongst
infants of diabetic mothers admitted into a special care baby unit in Port Harcourt ,
Nigeria. 2010;2–7.

59
62. Zhang Z. Model building strategy for logistic regression: Purposeful selection. Ann
Transl Med. 2016;4(6):4–10.

63. Mengesha HG, Wuneh AD, Lerebo WT, Tekle TH. Survival of neonates and predictors
of their mortality in Tigray region, Northern Ethiopia: Prospective cohort study. BMC
Pregnancy Childbirth [Internet]. 2016;16(1):1–13. Available from:
http://dx.doi.org/10.1186/s12884-016-0994-9

60
 APPENDIX
Appendix I Information Sheet
Title of the Research Project: Survival status and predictors of mortality among neonate
admitted to NICU from (2018-2020) at Dessie Referral Hospital, Northeast Ethiopia, 2021.

Name of Investigator: Mohammed Kebede (BSc in Nursing)

Name of the Organization: Addis Ababa University College of Health Science, School of
Nursing and Midwifery, Department of Pediatrics and Child Health.

Name of the Sponsor: Addis Ababa University.

Introduction: This information sheet is prepared for Dessie Referral Hospital administration
and NICU coordinating office. The aim of the form is to make the above-concerned office clear
about the purpose of research, data collection procedures and get permission to conduct the
research.

Purpose of the Research Project: To determine Survival status and predictors among neonate
admitted to NICU from 2018-2020 at Dessie Referral Hospital, Northeast Ethiopia, 2021.

Procedure: In order to achieve the above objective, information which is necessary for the study
was taken from neonatal medical records.

Risk and /or Discomfort: Since the information was collecting from medical charts, it will not
cause any hurt on the patients. The name or any other identifying informations were not
documented on the questionnaire and kept firmly private and in a safe dwelling. The information
retrieved were only used for the study purpose.

Benefits: The study have no direct profit for one whose document is involved in this study.
However, the indirect profit of the study for the participant in the program is obvious. The reason
is that if program developers are formulating predicted strategy there is a profit for clients in the
program of attaining proper care and treatment for clients survived and other newborns. In all,
the study work has a principal direct profit for health care organizers and administrators.

Confidentiality: To assure privacy the informations on the charts was collected by excluding
the name of the clients and the information collected from this study development was kept

61
private and stored in a file cabinet. In addition, it was not shown to anyone except the
investigator and it have been kept in a key and locked system with computer password.

Person to contact: This study development was revised and approved by Addis Ababa
University, College of Health Science, School of Nursing and Midwifery. If you have any
question you can contact any of the following individuals (Investigator and Advisors) and you
may ask at any the time you want.

Mohammed Kebede, Addis Ababa University, College of Health Science, Department of

Pediatrics and Child Health Nursing: principal investigator

Cell phone: +251-913803674, E-mail: mohama26.mk@gmail.com

Dr.Girum Sebsibe (MSC, PHD) Addis Ababa University, College of Health Science, School of
Nursing and Midwifery, Department of Pediatrics and Child Health Nursing: main Advisor.

Cell phone: +251-920856732

E-mail: girumseb@gmail.com

Mrs. Feven Mulugeta (BSC, MSc) Addis Ababa University, College of Health Science, School
of Nursing and Midwifery, Department of Pediatrics and Child Health Nursing: co- Advisor.

Mobile: +251910712396

E-mail: fevishome@gmail.com

Appendix II Checklist
This checklist is organized for the gathering of socio-demographic, maternal medical and
obstetrics and gynecology, neonatal medical and other key predictors and outcomes related data
that are significant for the valuation of survival status and predictors of mortality among
neonates admitted to NICU in Dessie Referral Hospital. All this information was retrieved from
the client's registration book and from an individual patient card without including the name of
the clients from [2018-2020]. This information was collected by health care providers (BSC
nurses) who are working in the NICU of the Hospital.

Contact information: Mohammed Kebede, Cell phone+251-913803674, Dr.Girum Sebsibe,

Cell phone: +251-920856732, Mrs. Feven Mulugeta, Cell phone: +251-910712396

62
Data collection date-------------------------month-------------------------Year---------------------

Name of the Hospital ----------------------------------------------------------------------------------

Name of data collector--------------------------- signature------------------------------------------

Name of supervisor---------------------------------signature-----------------------------------------

Code no---------------------------------------------------------------------------------------------------

Table 10: Data collection checklist for conducting of study on survival status and predictors of
mortality among neonates admitted to NICU in Dessie Referral Hospital, Northeast Ethiopia,
2020

Question for the mother

1. Sociodemographic characteristics
Question number Question Answer
101 Age 1. < 20 years
2. 20-35 years
3. > 35 years
102 Place of residence 1. Rural
2. Urban

2. Obstetric and gynecological related factors

201 Number of Parity ………………
202 Does the mother have 1. Yes
ANC follow up? 2. No
203 Was the current 1. Yes
pregnancy multiple 2. No
(twin)?
204 Which among the 1. PROM
following do you have 2. Preeclampsia
diagnosed? 3. Placenta previa
4. Prolonged labor
205 Mode of delivery 1. Spontaneous vaginal
delivery
2. Instrumental delivery
3. Cesarean section
206 Place of delivery 1. Hospital
2. Health center
3. Home
4. Other, specify……….
207 Delivery conducted by 1. Health professional

63
 2. Traditional birth
attendance
3. Other…….

1. Medical factor

301 Has she been diagnosed 1. Yes

with any medical 2. No
problems
302 If yes for question no 1. HIV
301, what was the 2. Hypertension
diagnosis 3. Anemia
4. DM
5. STI
6. Other, specify……

Questions for the neonate

1. Sociodemographic characteristics
Question Question Answer
number
101 ID no- ……….
102 Date of admission ……….
103 Age in day ……….
104 Sex 1. Male
2. Female
105 Gestational age at birth in weeks ………..
106 Weight in grams at birth ………..

2. Neonatal medical and other related factors

201 Mention the diagnosis of 1. Respiratory distress

neonate 2. Jaundice
3. Perinatal asphyxia
3. Hypothermia
4. Hyperthermia
5. Sepsis
6. Meconium aspiration
syndrome
7. Congenital malformation
8.Hypoglycemia
9. Other(specify

64
202 Bag and Mask Resuscitation at 1. Yes
birth 2. No
203 APGAR score of 1. First minute…….
2. Fifth minute………

204 Time of initiation of breast 1. Initiated within 1 hour

feeding 2. Not initiated within 1
hour
205 Length of hospital stay …………….
206 Date of discharge …………….
207 Time of neonatal death ……………..
208 Patient outcome 1. Died
2. Recovered
3. Referred
4. Left against medical
advice

65
Appendix III Approval Sheet
I, the undersigned MSc student, agree to accept responsibility for the scientific, ethical and
technical conduct of the research project entitled survival status and predictors of mortality
among neonates admitted to neonatal intensive care unit in Dessie referral hospital, Northeast
Ethiopia, 2020/2021, and provision of required progress reports to my advisors. All the
necessary advices and guidance will be asked and gained from my advisors.

Submitted by:

Mohammed Kebede (BSc) __________________ ________________

Name of student Signature Date

Approved by

ADVISORS:

Dr.Girum Sebsibe (MSC, PHD) ___________ _____________

Name of major advisor Signature Date

Mrs. Feven Mulugeta (BSc, MSc) _______________ __________________

Name of co -Advisor Signature Date

I, the undersigned, examiner have read, evaluate and attend proposal defense prepared by
Mohammed Kebede entitled with “survival status and predictors of mortality among neonates
admitted to neonatal intensive care unit in Dessie Referral Hospital, Northeast Ethiopia,
2020/2021.” This is to verify that his proposal has been accepted in partial fulfilment of the
requirements for the degree of masters in Pediatrics and child health nursing.

EXAMINER:

Rajalakshmi Somanathan (Ass. Professor, PHD): ______________ _______________

Name Signature Date

66