Page 58 - IJPS-7-2

International Journal of
Population Studies Child mortality by residence in Ethiopia

Table 1. Background characteristics of study participants in Second, Model I was fitted after including community-
Ethiopia, 2016 level predictors in the null model. Significant mortality
differentials were observed by place of residence and
Variable (N=8,162) Category N % regional category in Model I. Statistically significant
Regional category Emerging 3,703 45.47 (P < 0.001) lower risks of death were found among
Developed 3,706 45.41 children born in the urban areas relative to children in
Urban 753 9.23 the rural areas. Children from the emerging regions had
Place of residence Rural 6,680 81.8 higher risks of death compared with their counterparts in
Urban 1,482 18.2 the developed regions of the country.
Sex of household head Male 6,412 78.56 Third, Model II was fitted after including household-
Female 1,750 21.44 level predictors into the Model I. Similar to Model I,
Household size <6 3,633 44.51 children born in urban areas had statistically significant
(P < 0.001) lower risks of death than those children in
6+ 4,529 55.49
Combined wealth status Poor 4,473 54.8 the rural areas. Children living in the emerging regions
had statistically significant (P < 0.001) higher risks
Non-poor 3,689 45.2 of death as compared with their counterparts in the
Source of drinking water Improved 3,550 45.5 developed regions of the country. Children living within
Unimproved 4,612 56.5 small household size (<6 members) had a significantly
Type of sanitation facility Unimproved 6,782 83.09 (P < 0.001) higher risk of death as compared to those
Improved 1,380 16.91 from households of large size (six and above members).
Type of cooking fuel Solid fuel 7,731 94.72 Children from households with improved sanitation
Clean fuel 431 5.28 facilities had significantly lower (P < 0.001) risk of
Place of birth Home 5,770 70.7 death than their counterparts living in households with
unimproved sanitation facilities.
Health facility 2,392 29.3
Finally, the full model (Model III) with all proposed
Sex of child Male 3,910 47.9 explanatory variables including individual-level
Female 4,252 52.1 predictors was fitted to examine the effect of residential
Breastfeeding initiation (N=6911) Delayed 1,331 19.26. location on child mortality. Significant mortality
Immediately 5,580 80.74 differentials were observed at community-, household-,
Size of a child at birth Large 2,592 31.8 and individual-level attributes. In Model III, we found
Average 3,403 41.7 that children from the emerging regions had higher
Small 2,167 26.5 risks of death compared with their counterparts in the
Birth order First 1,609 19.7 developed regions of the country. Similarly, the lower
risks of death were found among children born in the
2 – 4 3,579 43.9 urban areas as compared with children in the rural areas
5 and above 2,974 36.4 as observed in Models I and II.
th
Age at child birth <18 459 5.62 As we found in Model II, children from the small
18 and above 7,703 94.38 household size (<6 members) had a significantly
Maternal education No education 5,399 66.2 (P < 0.001) higher risk of death as compared to those
Primary or 2,763 33.8 from households of large size (six and above members).
above
Likewise, children from households with improved
Total children ever born 1 – 4 4,665 57.16 sanitation facilities had significantly lower (P < 0.001)
5 and above 3,497 42.84 risk of death than their counterparts living in households
Religion Orthodox 2,342 28.7 with unimproved sanitation facilities. Children born with
Muslim 4,181 51.2 large and medium size at birth had a significantly (P <
Others 1,639 20.1 0.001) lower risk of death as compared to small size at
birth. This study also revealed that children born at health
facilities had a lower risk of death as compared to their
significant variation in child mortality across individual counterparts born at home. Children born to Orthodox
and household levels by place of residence justifying the Christianity follower women had a statistically significant
applicability of multilevel models for analysis. lower risk of death than children born from Muslim

Volume 7 Issue 2 (2021) 52 https://doi.org/10.36922/ijps.v7i2.392