Under-ve Mortality and Associated Factors among Children Born from 15-49 Years Old Mothers in Haramaya Town, Eastern Ethiopia

Background: The global under-ve mortality rate has dropped from almost 90 deaths per 1,000 live births in 1990 to 43 in 2015. The Ethiopian Mini Demographic Health Survey, 2019 shows 55 under-ve deaths per 1,000 live births. In the eastern part of Ethiopia, evidence from the Kersa Health and Demographic Surveillance System in Kersa district of East Hararghe Zone, Oromia Region suggested the decline of under-ve mortality rate from 131.8 per 1000 live births in 2008 to 77.4 per 1000 live births in 2013. The death rates still remain far from the Sustainable Development Goals’ target reduction to 25 or less per 1000 live births by 2030. However, the magnitude and determinants of under-ve mortality is not studied in Haramaya town. Objective: To assess the under-ve mortality rate and its associated factors among children born from 7 th August 2015- 6 th August 2020, in Haramaya town, east Ethiopia by 7 th –31 st August 2020. Methodology: Quantitative cross-sectional population-based study was conducted on 391 pairs of 15-49 years old mothers and their live-born under-ve children selected using systematic random sampling technique from Haramaya town to compare child mortality between <=24 and >=25 mother’s age groups. Data were collected using interview-based questionnaire; double entered into EpiData 3.1; and then exported to statistical package for social sciences program version 20.0 for analysis. Binary logistic regression analysis (p-value <0.20) was performed to examine crude association of predictors with under-ve mortality, and then multiple logistic regression analysis (p-value <0.05) to measure the statistical association. Results: The death of 28 out of 372 live births gave an under-ve mortality rate of 75 per 1000 live births. Children born in households with less than 6 members had 7. 98 times higher odds of dying than those born in households with at least 6 members (AOR =7.98, 95% CI =1.59-40.17). Those children who did not feed colostrum were associated with 17.45 times increased risk of under-ve deaths compared to colostrum-fed ones (AOR =17.45, 95% CI =6.54-46.55). The study suggests that 75 per 1000 live births die before celebrating their fth birth day.


Back ground
The child mortality rate (U5MR) has been widely used as an indicator of equality and human development (Alimohamadi et al., 2019). The U5MR is de ned as the probability (per 1,000 live births) that a child will die before reaching the age of ve if subject to current age-speci c mortality rates (Rostami et al., 2015). It is an important indicator that re ects the health of children and the development of the economy and culture of a country or region (Cao et al., 2017). It is one of the health indicators used by the World Health Organization (WHO) to assess a country's progress with improving the health of its citizen (Mdala and Mash, 2015). The mortality of children under 5 years of age forms Goal 3 of the Sustainable Development Goals (SDG 3) to reduce the U5MR to at least as low as 25 per 1000 live births by 2030 (United Nations, 2018). In 1990, the United Nations Development Program (UNDP) introduced the Human Development Index (HDI) which has three aspects: longevity, knowledge, and life standards respectively measured by life expectancy at birth, a combination of adults' literacy and enrollment rate, and Gross Domestic Product (GDP) per capita with the main aim of creating a more comprehensive measure of human development program (Alimohamadi et al., 2019).
Globally there has been an overall decrease in under-ve mortality since 1990 when many countries started working towards the fourth Millennium Development Goal (MDG), which aimed to decrease underve mortality by two-thirds by the year 2015 (Mdala and Mash, 2015). The number of under-ve deaths worldwide has declined from nearly 12.7 million in 1990 to 6.3 million in 2013. Consequently, the global U5MR has dropped from almost 90 deaths per 1,000 live births in 1990 to 46 in 2013 (Kipp et al., 2016).
From 1990 to 2015, the global rate of U5MR declined by more than half and reached 43 per 1000 live births which was less than MDG 4 and far from the target Sustainable Development Goals (SDGs) to end preventable deaths of children under 5 years by 2030 (Alimohamadi et al., 2019). However, this progress is not equally distributed at national and sub-national levels (Rostami et al., 2015).
From an ecological study in West and South Asian countries, the highest rate of decrease was related to India (167.6 to 59.9 in 100 live births from 1980-2010) and Iran (109.9 to 19.2 in 100 live births from 1980-2010) (Alimohamadi et al., 2019). There are few studies on sub-national child mortality (CM) in Iran, but a published report by the Iranian Ministry of Health showed that in 2010, more than 91% of mortality in the neonatal period and more than 63% of mortality 1 to 59 months after birth occurred in Iranian hospitals (Rostami et al., 2015). Deaths in neonates accounted for more than 50% of all deaths among under-5 children during the 24 consecutive years under investigation in Beijing (Cao et al., 2017).
African children are nearly 16 times more likely to die under age ve than children from high-income nations, though progress in Africa is being made (Kipp et al., 2016). U5MR in the sub-Saharan region has remained unabated (Yaya et al., 2018). Many countries, especially those from sub-Saharan Africa, were unable to reach the Millennium Development Goal for under-5 mortality reduction by 2015 (Mdala and Mash, 2015). The current U5MR (per 1,000 live births) was; 133 in the Republic of Chad, 104 in the Democratic Republic of Congo, 95 in Mali, 127 in Niger, and 69 in Zimbabwe (Yaya et al., 2018).
In Ethiopia, it declined substantially in most of the Health and Demographic Surveillance System (HDSS) sites during the last 10-15 years (Amare et al., 2016). For instance, evidence from the Kersa HDSS in Kersa district of East Hararghe Zone, Oromia Region, Eastern Ethiopia suggested the decline of U5MR from 131.8 per 1000 live births in 2008 to 77.4 per 1000 live births in 2013 . The prevalence rate of under-ve mortality in Ethiopia was 67 per 1000 live births from Ethiopian Demographic Health Survey (EDHS) 2016 (Central Statistical Agency andICF, 2016). This represents a 60% decrease in under-ve mortality over the period of 16 years since 2000 when there were 166 deaths per 1000 live births (Shewayiref and Setegn, 2020).
It seems that improving maternal education level, increasing age at marriage, and the birth gap between two births has an important role in decreasing U5MR in India (Alimohamadi et al., 2019). For girls there are also major bene ts for the next generation: half of the reduction in under-5 mortality achieved in the last 30 years may be attributable to increased maternal education (Glynn et al., 2018).

Statement of the problem
In 2015, it was estimated that 5.9 million children aged under 5 years died (Alimohamadi et al., 2019).
Evidence from Ethiopian Mini Demographic Health Survey (EMDHS), 2019 shows 55 under-ve deaths per 1,000 live births (Ethiopian Public Health Institute and ICF, 2019).
The younger the mother is, the higher is the risk of complications both for her and the child, and childbearing before age 16 is of particular concern (Sandoy et al., 2016). However, evidence from Ethiopian Demographic and Health Survey, 2016 showed that about 40% of women had given their rst birth before age of 17 years old (Berhanu, 2019). The World Health Organization estimates that infant deaths are 50% higher amongst babies born to mothers aged below 20 years (Johnson et al., 2019). Infant mortality and morbidity, preterm birth, low birth weight, and asphyxia are also signi cantly higher among infants born to women under age 20 compared to older women (Luseno et al., 2016). A child born to a teen mother is twice as likely to die before the rst birthday as compared with the child of a woman in her 20s (Kamal and Hassan, 2015). The majority of child (6.36 %) death occurred due to the mothers' age at birth is less than 18 years in Ethiopia (Dereje et al., 2018). These high rates of mortality and morbidity may be, in part, due to immature physical structure, low utilization rates of maternal and child health services and low education levels among adolescents (Luseno et al., 2016). A range of demographic and socioeconomic factors are known to account for enormous disparities in the uptake of maternal health care in low-and middle-income countries (Sekine and Carter, 2019).
Globally almost 40% of all under-ve deaths are due to preventable or treatable infectious causes, particularly bacterial pneumonia and diarrhea diseases, as well as birth complications and malnutrition (Mdala and Mash, 2015). Several studies have shown that the Mother To Child Transmission (PMTCT) for HIV and malnutrition have an important in uence on childhood mortality (Mdala and Mash, 2015). There is also high neonatal mortality in many poor countries due to premature birth, birth asphyxia, and sepsis whereas the deaths in older children after the perinatal period have remained high due to low coverage of pneumococcal vaccine, poor sanitation and water supply, and poor food security in low-income countries, as well as displacement of populations in countries with war or con ict (Mdala and Mash, 2015). The U5MR increased by 7.20% from 2013 to 2015 in Beijing, China where birth asphyxia, congenital heart disease, preterm/low birth weight, and other congenital abnormalities comprised the top ve causes of death (Cao et al., 2017). The associated risk factors of under-ve mortality in rural settings of Ethiopia include differences in regions, educational attainment, born in twin, place of delivery, availability of occupation of parents, age of mothers at rst birth, breastfeeding status, birth order, religious belief, use of a contraceptive method, child vaccination, and family size (Shewayiref and Setegn, 2020). In the study in Ethiopian Somali Regional State, family size, preceding birth interval, birth order, breastfeeding status, and source of drinking water were signi cant determinants of under-ve mortality in the region (Solomon et al., 2017).
Generally, it was observed that U5MR and maternal mortality rate had a decreasing trend from 1980-2010 (Alimohamadi et al., 2019). One of the most important causes of this decrease was the improvement of HDI; thus, by increasing educational levels, income per capita, and life expectancy we might observe a substantial reduction in maternal and child mortality rates (Alimohamadi et al., 2019). In a French study, it was seen that U5MR decreased from 1990 to 2011 in all low-and middle-income countries and this decrease was in uenced by factors such as improvement of health care and economic status (Alimohamadi et al., 2019). The reduction in mortality was consistent with changes in the proximate determinants of child survival in Bangladesh which included maternal factors, environmental contamination, nutrient de ciency, personal illness control, and injury (Chowdhury et al., 2017). Immunization and management of childhood illnesses are two important personal illness control measures, which have contributed enormously in reducing under-ve mortality (Chowdhury et al., 2017). Health and population programs have been effective in increasing immunization coverage, use of ORS (Oral Rehydration Solution) for managing diarrheal diseases, and increasing contraceptive use (Chowdhury et al., 2017). India achieved an impressive decline in under-ve mortality; from 167.5 in 1980 to 45.2 per 1000 live births in 2015 and reduced diarrhea mortality was an important contributing factor (Choudhary et al., 2019). ORS use and reduction in stunting were the two key interventions, each accounting for around 32% of the lives saved during this period in the country (Choudhary et al., 2019). Steadily increased contraceptive use during the last couple of decades from 40% in 1990 to 62% in 2014 has reduced the risk of CM associated with higher-order births, short birth intervals and large family size, which might have contributed to reducing overall CM (Chowdhury et al., 2017).
The under-ve mortality rates, although reduced from 67 per 1000 live births evidence from EDHS, 2016 (Central Statistical Agency andICF, 2016) to 55 per 1000 live births from EMDHS, 2019 (Ethiopian Public Health Institute and ICF, 2019), remain far from the SDGs target. The study analyzed data from the EDHS, 2016 was based on secondary data the major limitation was that some important determinant factors of under-ve mortalities like previous death of a child and maternal TT immunization were missed due to high missing values in the data. The other weakness of the study was the mother's recall of events that took place for the past ve years preceding the survey which was subject to recall bias (Berhanu, 2019).
Knowledge about the current under-ve mortality and its determinant factors plays a great role in adapting different strategies and methods to achieve the desired goal, but there is little known about it. This knowledge is not only useful for enabling us to achieve SDGs desired target of reduction in under-ve mortality, but also with proper measures makes it possible to reduce the aftermath consequences for the country. Hence further investigation had to be conducted including variables not included in this study and to use spatial models to account spatial variation of experiencing under-ve mortality since regional variations are signi cant (Berhanu, 2019). However, the magnitude and determinants of under-ve mortality is not studied in Haramaya town.
Therefore, this study aimed to examine the under-ve mortality rate and associated factors among those children born ve years before the study in the town.  (Chowdhury et al., 2017). In a study conducted in 2015, the mean years of schooling showed a signi cant correlation with the decrease of mortality rates meaning as education level increased, health literacy of women rose as well which led to better health, better care of their children, and more attention of women to the health of their family, and thus reducing the mortality and morbidity in West and South Asian countries (Alimohamadi et al., 2019). Children of mothers with no schooling had 1.88 times higher mortality than those whose mothers had six or more years of schooling in Bangladesh (Chowdhury et al., 2017). The U5MR was found to be signi cantly lower among children born to parents who attended at least a high school education in Bhutan (Dendup et al., 2018). The odds of UFM were signi cantly lower, 62% and 77%, among children whose fathers had a high school and tertiary level of education respectively and 69% lower among whose household head had received high school education (Dendup et al., 2018).
For instance, in South Africa, falling behind in school was the strongest risk factor for giving birth within the following two years (Glynn et al., 2018). The study results in Nigeria showed that the risk of childhood mortality is 26.7%, 39.7 and 45.9% lower among the mothers having primary, secondary and tertiary education respectively than those with no formal education (Yaya et al., 2017). One of the pathways by which mothers' education affects child survival is through improved child care (Yohannes et al., 2017). This is probably because those with little education before marriage may be unable or poorly equipped to take appropriate reproductive health decisions that are important in reducing pregnancy-related complications (Adedokun et al., 2017). Increased schooling has been associated with better health of women and their children in Zambia (Sandoy et al., 2016). An analysis of the 2011 EDHS data reveals that the risk of dying for a child born to an uneducated mother was 2.13 times higher compared to a child whose mother had primary and higher education. Also in a Gilgel Gibe eld research of determinants of UFM, children born to mothers whose educational level was below elementary were 14 times more likely to die compared to children whose mothers' education is above elementary (Yohannes et al., 2017). Similarly, evidence from EDHS 2016 data analysis suggests that deaths among under-ve children differed signi cantly with the level of mothers' education, with those of relatively higher education, having a lower chance of experiencing under-ve deaths (P-value < 0.001) (Berhanu, 2019). Children born to a mother with no education at all were associated with a 2.61 times increased risk of under-ve deaths compared to being born to a mother with higher education (OR=2.610, 95% CI: 1.598, 4.265); children from mother with only primary education were 2.27 times more likely to at the risk of under-ve mortality (OR=2.271, 95% CI: 1.398, 3.687), while children born to mother with secondary education were 2.163 times more likely to die before celebrating their fth birthday (OR=2.163, 95% CI: 1.184, 3.534) compared to being born to mother with higher education, keeping all other covariates constant (Berhanu, 2019). Evidence from Kersa HDSS in Kersa district of Eastern Hararghe, Oromia, Ethiopia also shows that maternal educational status was signi cantly associated with under-ve children's mortality OR 1.31 (1.13, 1.49) (Melkamu et al., 2015).

2. Income
Children from low asset category households had on average 1.17 times higher mortality rate than those from high asset category households in Bangladesh (Chowdhury et al., 2017). The odds of UFM reduced signi cantly with increasing wealth quintiles and were lower among those born to working mothers (Dendup et al., 2018). The risk of childhood mortality was signi cantly lower in the middle, richer and richest (11.1%, 37.5 and 49%) economic quintiles respectively when compared to the risk of childhood mortality with a female spouse who is poorest in Nigeria (Yaya et al., 2017). Low empowerment, indicated by low income and low decision-making power is signi cantly associated with the likelihood of experiencing pregnancy complications in the study area as those with low empowerment are (OR 3.962, CI 0.937-1.083, P<0.005) more likely to experience complications when compared with other categories (Adedokun et al., 2017). Evidence from Kersa HDSS in Kersa district of Eastern Hararghe, Oromia, Ethiopia also shows that low household wealth index was signi cantly associated with under-ve children's mortality OR 1.26 (1.10, 1.43) (Melkamu et al., 2015).

3. Household size
Children born in households with at least 6 members experienced lower mortality rates. The likelihood of mortality was 66% lower (95% CI: 0.21-0.55) among children born in households with > 5 members than those born in households with less than 6 members (Dendup et al., 2018). The possible in uence of interaction between household size and other variables (wealth index, mother's working status, mother's education level, and place of residence) on UFM was found to be not signi cant, and is thus not reported. All variance in ation factors (VIFs) were < 10, suggesting that multicollinearity was not a concern in the regression analysis (Dendup et al., 2018). A study in an Ethiopian Somali regional state suggested that family size was a signi cant determinant of under-ve mortality in the region (Solomon et al., 2017). According to the EDHS 2011 data analysis, the mortality risk of children increases as the size of the family increases in the region. The risk of dying for a child born in family of size 4-6 is higher by 1.879 times relative to those born in family with size of 1-3 (reference category). Children born in a family with the size of seven and above have a signi cantly higher hazard rate than children born in a family with the size of 1-3, i.e., children born in a family of size seven and above have a 2.164 (HR=2.164, 95% CI: 1.987, 8.215) times higher risk of death as compared with children in the reference category (Solomon et al., 2017).

4. Occupation of mother
The employment status of mothers and husbands were identi ed as signi cantly associated factors with under-ve deaths (P-value < 0.05). The odds of under-ve mortalities were 18.4% (OR=0.816, 95% CI: .666, .999) lower among women who were to working at all compared to those who were skilled or manual worker. The probability of child mortality, under-ve, was 32% (OR=0.679, 95% CI:.533, .864) and less likely to occur among women with an unemployed husband and 22% (OR=.783, 95% CI: .618, .991) less likely to occur among women whose husbands were professionally employed compared with those who were working as sales and others (Berhanu, 2019).

5. Availability of electricity
The likelihood of UFM was also signi cantly higher among children born in households without electricity, in the eastern and central regions, and those living in rural areas (Dendup et al., 2018). The odds of UFM were signi cantly higher among children born in households without electricity (AOR = 1.81, p = 0.026) and those born in the central (AOR = 1.72, p = 0.025) and eastern (AOR = 2.09, p < 0.001) regions (Dendup et al., 2018

Place of delivery
Nigeria observed that high stillbirth and early neonatal mortality rate have been long associated with unattended deliveries compared with hospital-based deliveries (Adedokun et al., 2017). Evidence from Ethiopian Demographic and Health Survey, 2016 showed that the percentage of under-ve mortalities is higher among home deliveries, which is about 27.6% compared to 24.7% among health facility deliveries (Berhanu, 2019). Evidence from Kersa HDSS in Kersa district of Eastern Hararghe, Oromia, Ethiopia also suggested that place of delivery was signi cantly associated with under-ve children's mortality OR 1.26 (1.10, 1.43) 1.016 (1.013, 1.12) (Melkamu et al., 2015).

Colostrum feeding
Colostrum feeding practice after birth was associated with UFM, whereas the rest of the health-related variables were not found to be signi cant (COR =2.26, p = 0.042) (Dendup et al., 2018). According to a study conducted in Afambo district of Afar Regional State, the chi-square test showed that colostrum feeding was signi cantly associated (p =0.035, p =0.006 and p =0.001) with the three indicators of child under-nutrition (stunting, wasting and underweight) respectively (Misgan et al., 2016) (Dendup et al., 2018). Multivariate analysis showed that having >1 under-5 child death was associated with maternal age at delivery (<18 or >35) of mothers in a study conducted in Iran (Anafcheh et al., 2018). Compared to those born to younger (≤25 years) mothers, children born to mothers aged 36-40 years, 41-45 years, and more than 45 years had signi cantly lower odds of UFM (Dendup et al., 2018). Children born from mothers whose age less than or equal to 24 have a signi cantly lower risk of under-ve mortalities (OR=0.295, 95% CI: .227,.383) compared to those born from mothers whose age is between 45 and 49 in Ethiopia (Berhanu, 2019).

Mother's age at rst marriage
The U5MR was found to be higher among children whose mothers were older than 45 years, who were younger than 16 years when they rst married (Dendup et al., 2018).

Mother's age at rst birth
Maternal age at rst birth was identi ed as a strong predictor of under-ve mortalities in both bivariate and multivariate analysis after controlling for the effects of other covariates (Berhanu, 2019). A systematic review revealed that a mother's age at rst birth is negatively correlated with infant mortality (IM) in Ethiopia. Its effect (except children born to mothers older than 20 years of age at rst births) has a signi cant impact on CM (Yohannes et al., 2017). Similarly, the age of the mother at rst birth was signi cantly associated with under-ve mortalities in Ethiopia (P-value < 0.0001). The risk of under-ve mortality was about 55.6% higher for births to mother give birth at earlier age 11 to 17 years compared with births to mothers 25 and higher years old (OR=1.556, 95% CI: 1.243, 1.949) (Berhanu, 2019).

Sex of the child
In Bangladesh, the present analysis examined under-ve mortality in which the neonatal deaths formed the major part due to rapid decline in childhood mortality during age 1-4 years of life and the higher under-ve mortality for male than female was due to lower female mortality during the neonatal period (Chowdhury et al., 2017). The under-ve mortality rate steadily declined over the years from 128/1000 in 1994 to 48 in 2014 and females had 8% lower mortality rates than males (Chowdhury et al., 2017). The U5MR was found to be higher among male children in Bhutan. Boys were 1.38 times more likely to die than girls (p = 0.055) (Dendup et al., 2018). The odds of under-ve death were also higher for male children (AOR=1.30, CI= [1.07, 1.57], P<0.008) compared to females in Ethiopia (Chaltu et al., 2019).

Number of births
Several children is also another signi cant factor identi ed in the analysis, it was revealed that everexperienced complications may likely increase with the number of children, an interesting nding in this study, particularly with low contraceptive use (Adedokun et al., 2017). Children born to mothers who gave birth to more than 2 children had signi cantly higher odds of dying. The odds of death among children born to mothers who gave birth to 3-4 or more than 4 children were signi cantly higher than those born to mothers with less than 3 children (Dendup et al., 2018). The U5MR was found to be higher among children whose mothers had more than 4 births in Bhutan (Dendup et al., 2018 EDHSs; particularly the 2nd and 3rd birth order were dominant determinants. The estimated hazard ratios of mortality were higher for rst birth orders compared to second and third ones (Yohannes et al., 2017). The study result in Ethiopia indicates hazard ratio of children who had birth order fth and above was 1.683 [95% CI: 1.190,2.380]. This means children born at the fth and above were 1.683 times more likely to die than children born at 1st-2nd order (Dereje et al., 2018). A study in an Ethiopian Somali regional state also suggested that birth order was a signi cant determinant of under-ve mortality in the region (Solomon et al., 2017). According to this study, higher birth orders (>4) have the highest mortality risk. Children with these characteristics are 2.067 times more likely to die in age less than 5 relative to the reference group births of order one (HR=2.067, 95% CI: 1.098 to 8.256). Children of order two through four are dying at a rate 23.6% higher than a child of order one (HR=1.236, 95% CI: 1.031 to 12.199). The con dence intervals for higher birth order and birth order two through four indicate that the rate could actually be as high as 8.256 and 12.199 and as low as 1.098 and 1.031, respectively (Solomon et al., 2017 A study in an Ethiopian Somali regional state suggested that the source of drinking water was a signi cant determinant of under-ve mortality in the region (Solomon et al., 2017). The risk of dying for a child born in a family without access to pipe drinking water is higher by 76% relative to those born in a family with access to pipe drinking water. The 95% con dence interval (1. 421, 9.373) implies that the risk of death of children whose source of water is not pipe water is 1.421 as low and 9.373 as high as those in the reference group (Solomon et al., 2017).

Safe sanitation facilities
The likelihood of UFM was 1.49 times higher for those children born in households without safe sanitation facilities than their counterparts (p = 0.012) (Dendup et al., 2018).

Solid fuel use
Furthermore, the U5MR was signi cantly higher among children born in households that used solid fuel.
The likelihood of UFM was 2.18 and 1.95 times signi cantly higher among children born in households without safe sanitation and those that used solid fuel for cooking respectively (Dendup et al., 2018

Study Area and Period
Haramaya town is situated in East Hararghe Zone, Oromia Regional State, Ethiopia. It is located 21 km Northwest of Harar town and 505 km East of Addis-Ababa, the capital city of Ethiopia. The elevation in the site ranges from 2018 meters to 2422 meters above sea level (m. a. s. l). The watershed lies at a UTM (Universal Transverse Mercator) coordinate of 171212 East and 1040190 North. It is characterized by a "Woina-Dega" agro-climatic zone that receives a mean annual rainfall of 775.9 mm. The monthly rainfall in the site is more than 100 mm from April to September, except June 48.4 mm. The wettest month is August, 151.9 mm. The daily temperature in the site ranges from 10 o C -25 o C. The livelihood in the area is based on agriculture. Therefore, different agricultural practices are practiced within the Lake Haramaya watershed (Haile and Ararso, 2016). The town has one primary hospital and is divided into 3 administrative kebeles and 24 sub-kebeles where 9, 7 and 8 of them belong to kebeles 01, 02 and 03 respectively. The estimated population of the town was 50960 of which 8373 were expected to be children under-ve years of age. Of the population, 19030 (3127 under-ve), 18214 (2993 under-ve) and 13716 (2253 under-ve) were expected to live in kebeles 01, 02 and 03 respectively. The childhood mortality status in the town is yet unknown. The study was conducted starting from August 07-31/2020.

Study Design
A community-based cross-sectional study design was selected.

Source Population
All pairs of 15-49 years old mothers lived and their children born alive in Haramaya town in the past ve years.

Study Population
All 15-49 years old mother-live-born under-ve child pairs who were living in randomly selected subkebeles of Haramaya town during the study.

Inclusion criteria
All 15-49 years old mothers who had had live births in the past ve years were included.

Exclusion criteria
Mothers who either had had multiple births like a twin or couldn't respond due to serious illness were excluded. Multiple births were excluded because of the known higher risk of neonatal mortality due to pregnancy complications and preterm birth amongst multiple births compared to singleton births to avoid potential confounding effects (Dendup et al., 2018, Ghimire et al., 2019.

Sample size determination
The sample size calculated using the single population proportion formula; n= (z α/2 ) 2 π(1-π)/d 2 for speci c objective one (to determine the under-ve mortality rate among the last children born in the past ve years in Haramaya town) is as follows where n is the sample size, π is population proportion and d is precision required in % = (1.96) 2 *0.067(1-0.067) /0.05 2 = 3.8416*0.067(0.933) /0.0025 = 0.24/0.0025 =96 Furthermore, the sample size was also calculated with the double population proportion formula; for speci c objective two of this study which was to identify the factors associated with under-ve mortality among the last children born in the past ve years in Haramaya town as indicated in table1 below using the following assumptions; con dence level of 95% and 5% signi cance level (Berhanu, 2019 Since the sample size calculated for speci c objective two accommodated the largest sample size, it was (237 x 1.5) + 10% = 391. Therefore, the minimum total sample size considered to undertake the study was 391 after considering design effect of 1.5 and non response rate of 10%.
3.6. Sampling procedure/technique Participants were selected using a two-stage sampling among 638 mothers of 15-49 years old age identi ed before the study whose births had been registered by the vital events registration agency. A simple random sampling method was used to select sub-kebeles from all the 3 kebeles of Haramaya town proportional to their population size. Accordingly, 3 of the 9 sub-kebeles of kebele 01 (sub-kebeles 1, 4 and 7); 2 of the 7 of 02 (3 and 6) and 2 of the 8 of 03 (2 and 7) were selected. Then systematic random sampling to select participants independently from the selected sub-kebeles proportional to their respective number of mothers of under-ve children identi ed. Since the sampling interval was 1 for all sub-kebeles, the sample was taken starting from the rst cases. Finally, the study was conducted on 391 participants as shown in gure 3 below.

Data Collection Tool
The data collection tool consisted of socioeconomic, socio-demographic, environmental, and healthrelated characteristics that would affect under-ve mortality. Questionnaire for primary data which included 17 structured and open-ended questions was adopted from different literatures put in its review as socioeconomic, bio-demographic, environmental, and health-related factors (Dendup et al., 2018, Ghimire et al., 2019, (Biniyam et al., 2018, Berhanu, 2019, Chaltu et al., 2019, Solomon et al., 2017 in English language rst and then translated into Afan Oromo by language experts and again the Afan Oromo version translated back to English to make it consistent. Data were collected using the Afan Oromo structured and open-ended questions.

Data Collectors
Three nurses who had previous experience in data collection were selected among my co-workers and given training. Then they collected data according to the schedule already set.

Data Collection procedures
Data were collected home-to-home using interview-based questionnaire from fteen study participants daily for the rst twenty two days, and then from fourteen participants daily for the last three days based on time availed for data collection. The data collection process was supervised and continuously followed up by the principal investigator throughout the data collection period.  (Biniyam et al., 2018) Safe drinking water: Household has access to a protected drinking water source (yes, no); piped tap, protected well and protected spring are safe water sources whereas surface water like a river, pond, etc.
are unsafe Safe sanitation facilities: Household has access to improved sanitation facilities (yes, no); water-ush toilet, ventilated improved pit latrine and traditional pit latrine are safe sanitation facilities whereas communal latrine, eld or anywhere are unsafe Use of solid fuel: Use charcoal, wood or kerosene for cooking (yes, no) Maternal TT immunization: Tetanus toxoid injection status of mothers (one or no TT, two or more TT) Contraceptive use: Use of contraceptives in the mothers' experience (ever used, never used) Place of delivery: Place of delivery for the last birth under ve years (health facility, home) Colostrum feeding: Feeding colostrum soon after birth to the last child under ve years (yes, no); colostrum is secreted with its potential bene ts within 1 hour of birth (Gebretsadkan et al., 2020) 3.10. Data quality control To maintain the quality of the data, data collectors were trained in data collection procedures. The questionnaire was carefully designed and prepared in English language rst and then translated into Afan Oromo by language experts and again the Afan Oromo version was translated back to English to make it consistent. Finally, the Afan Oromo version was used to collect data. Before actual data collection time, the questionnaire (tool) was pretested for validity and reliability in 10% of the sample in Aweday town. Data quality was also controlled by daily cleaning of data.

Methods of data analysis
Data were double entered into EpiData version 3.1; and then, cleared, coded, processed and analyzed using computer, statistical package for social sciences (SPSS) version 20.0. To descriptive statistics; frequencies and percentages were used. The U5MR was calculated as 'deaths per 1000 live births'. Multicollinearity checked among the independent variables was not there. Binary logistic regression analysis (Odds ratio, 80% CI and P-value 0.20) was performed to examine the crude association of predictors with under-ve mortality, and then multiple logistic regressions (Odds ratio, 95% CI and P-value 0.05) to measure the statistical association. In the multivariate analysis, backward logistic regression method was used with the Hosmer and Lomeshow googness-of-t test, and the value of 0.431 was insigni cant which means the nal model was correct. Finally, the results were presented in texts, tables, and gures.

Ethical considerations
The study was conducted after the proposal was approved by HU, CHMS IHRERC (Haramaya University,

College of Health and Medical Sciences Institutional Health Research Ethics Review Committee).
Informed, voluntary, written and signed consent was taken from each mother or husband in case of a minor. Con ict of interest was controlled. Privacy and con dentiality of participants were assured. The selection of the mothers was also equitable and fair. There was no risk of participating in this study than taking few minutes from their time. The ndings from this research revealed important information for the local health planners even though there was no direct payment for participating in the study. They also had the right to participate or not and even could withdraw at any time once decided to participate.

Socio demographic characteristics
Out of the calculated sample of 391 mothers included in this study who had had singleton live births from 7 th August 2015 to 6 th August 2020, 372 (95.14%) responded to the home-to-home interview. The median of current mothers' age was 24.5 with a standard deviation of +/-5.81824. The range of mothers' age was 15 to 45. The majority of the births (52.4%) were males; 64.2% in households with not more than 5 members. Of the mothers, 86.0% fed their child colostrum (Table 2).

2. Under-ve Mortality Rate
Out of 372 singleton live births from 7 th August 2015 to 6 th August 2020, 28 (7.5%) died. This gave a weighted U5MR of 75 per 1000 live births. Children born in households with less than 6 members experienced higher mortality rates. Signi cantly higher U5MR was also recorded among children who did not feed colostrum (p <0.001).  (-57.0, 307.0) a Weighted under-five mortality rates as 'deaths per 1000 live births'; b Confidence interval; TT = tetanus toxoid

Factors Associated with Under-ve Mortality
In bivariate analysis (p <0.20), households earn less than 1700 ETB per month, households with less than 6 family members, mothers working as house wives, mothers who had never used any contraceptive, children not colostrum-fed, mothers aged <=24 years, and mothers who had more than 4 births were found to be candidates for multivariate analysis (Table 3). In multivariate analysis, household size, and colostrum feeding status were signi cant factors associated with under-ve mortality of children (Table 4). Children born in households with less than 6 members had 7.98 times higher odds of dying than those born in households with at least 6 members (AOR =7.98, 95% CI =1.59-40.17, P <0.05). Those children who did not feed colostrum were associated with 17.45 times increased risk of under-ve deaths compared to colostrums-fed ones (AOR =17.45, 95% CI =6.54-46.55, P <0.001). The wide con dence intervals might be attributable to the small sample size of the study.
In this study, household size was signi cantly associated with under-ve mortality. This study found that household size was inversely associated with under-ve mortality consistent with the nding from Bhutan (Dendup et al., 2018). The ndings support the view that larger households may have better resources such as more experienced child care providers and more working-age adults contributing to the household income. In contrast, another nding in the Ethiopian Somali Regional State suggested that the mortality risk of children increases as the size of the family increases (Solomon et al., 2017). To this view, larger family size may indicate more children, leading to intra-sibling competition for limited resources and inadequate attention and care to children heightening their mortality risk. However, the nding in this study suggests that larger family size is protective against UFM in Haramaya town, possibly through more people to care for the child and resources.
The other prominent factor associated with under-ve mortality in this study was colostrum feeding practice after birth. The death rate was signi cantly higher among children not colostrum-fed in line with the study nding from Bhutan (Dendup et al., 2018). Colostrum is the rst milk containing proteins, vitamin A and maternal antibodies. A study conducted in the Afambo district of Afar Regional State showed that colostrum feeding was signi cantly associated with stunting, wasting, and underweight (Misgan et al., 2016). The rest of the variables were not signi cantly associated with UFM in this study. For instance, there was no association between UFM and mother's age unlike the study result in other parts of Ethiopia which stated that it was signi cantly lower among children born from mothers whose age less than or equal to 24 compared to those born from the reference group (Berhanu, 2019).
Mother's education was also not associated with UFM in Haramaya town. A similar report was obtained in a study from Ethiopian Somali Regional State (Solomon et al., 2017) and also in other parts of Ethiopia (Chaltu et al., 2019). The ndings of this study suggest that strengthening health education particularly on exclusive breast feeding practice can improve child survival.

Strengths and Limitations of the Study
This study is a rst of its kind to be undertaken in Haramaya town on under-ve mortality. Its high response rate, more than 95%, makes it generalizable to the target population. The use of valid survey methods lends to the study's credibility. Additionally, the use of an appropriate theoretical framework relevant to developing countries to assess under-ve mortality also makes this study strong. The most recent singleton live births ve years before the survey were considered for analyses to reduce maternal recall bias.
However, this study also had the following limitations. Firstly, the cross-sectional design of the study prevents drawing causal inferences on the associations revealed. The study design might have introduced recall bias and social desirability biases because of the nature of self-reported data in this study. The sample size used is small relative to that of previous studies. The survey only interviewed surviving mothers, and hence could have led to the underestimation of U5MR and the effect of various factors. Another drawback is that this research was unable to access the age interval where most deaths occurred and it also could not determine the causes of death. Besides, as the study was conducted in a single town, the results might not be representative of the country.

Conclusion
The study suggests that 75 per 1000 live births die before celebrating their fth birth day in Haramaya town. Household size and colostrum feeding status are the factors associated with under-ve mortality of children in the town.

RECOMMENDATIONS
Based on the nding of this study, the Haramaya town health o ce should disseminate information to the mothers about the bene ts of feeding their children colostrum and promote the practice. The o ce, in collaboration with other relevant sectors and programs, should do its best in strengthening child care services.
East Hararghe zonal health department should take part in the allocation of su cient budget for all these activities including research. It should also closely monitor and evaluate changes that will be there.
In addition, further research at the regional, as well as national level, will enable us to understand the magnitude and associated factors of children's mortality under-ve using a larger sample size and including variables not included in this study like mother's age at rst birth. Sampling procedure of selecting sub-kebeles and study participants in Haramaya town

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.