59 / 89
МОНГОЛЫН ХҮН АМЫН СЭТГҮҮЛ Дугаар (367) 20, 2011
58
Another factor that had statistically significant
relationship with child mortality was child’s gender
(p≤0.05). Male children died more than female
children with 0.76 per cent higher.
Child’s weight had quite strong statistically
significant relationship with child mortality
(p≤0.001). Children who weighed less than 2,500
gm at birth tended to die with 3.85 per cent more
than children who weighed more than 2,500
gm.
Another factor that had quite strong statistically
significant relationship with child mortality was
source of drinking water (p≤0.01). Households who
got their water from central or local pipe had child
mortality with 0.35 per cent less than the household
who got water from well. However, the percentage
of experiencing child mortality among households
who got water from central or local pipe was 1.84
per cent less than that of the household who got
water from spring/mineral/rain/snow/river.
Antenatal care had statistically significant
relationship with child mortality (p≤0.05). Mothers
who had less than four antenatal care visits tended
to experience child mortality with 1.65 per cent
more than that of those who had more than four
antenatal care visits.
Multivariate analysis
For the first model (Table 2), socio-demographic
variables were put in the model, and binary logistic
regression was run to examine the odds of having child
mortality by independent variables. Interestingly, place
of residence as regions had no statistically significant
effect on child mortality regardless of socio-economic
disparities and shortage of health care facilities in rural
areas. In Mongolia, all regions except for Ulaanbaatar
the capital city are considered as rural areas. However,
mothers lived in Eastern region were more likely to
experience child mortality 15% more than that of the
mothers lived in Ulaanbaatar the capital city. The odds
ratio of having child mortality in Western region was
1.78 times more, whilst the odds of child mortality in
Southern and Central regions were 1.1 and 1.0 times
more than that of the mothers in Ulaanbaatar. As for
ages, mothers aged 30 years old or more were 79 per
cent less likely to experience child mortality than those
who aged less than 20 years old at p≤0.01 level. Mothers
who aged between 20 and 29 had no statistically
significant impact on child mortality; but, the odds
of experiencing child mortality for this age group of
mothers comparing to the age group of less than 20 years
old were 55 per cent lower. Neither monthly income
per person nor mother’s education had a statistically
significant relationship with child mortality.
Table 2:
Odds ratios and regression coefficients for child mortality by selected characteristics
Characteristics
Model 1
Model 2
OR
B
OR
B
Place of residence
(ref=Ulaanbaatar)
East
0.85
-0.15
0.79
-0.23
West
1.78
0.58
0.94
-0.05
South
1.10
0.09
1.43
0.35
Central
1.04
0.03
0.81
-0.20
Mother’s age
(ref=less than 20 years)
20-29 years
0.45
-0.79
0.46
-0.77
30 and more years
0.21**
-1.53
0.19**
-1.61
Mother’s education
(ref=uncompleted secondary)
Complete secondary
1.08
0.07
1.19
0.17
Tertiary
1.11
0.11
1.16
0.14
Monthly income per person (
₮
)
(ref=less than 102,535)
More than 102,535
0.48
-0.72
0.65
-0.42
Ever breastfed
(ref=no)
Yes
------
------
0.04***
-3.22
Birth order
(ref=1 & 4 and above)
2-3
0.74
-0.30
Child gender
(ref=male
Female
0.52*
-0.65
Child weight
(ref=less than 2,500 gm)
More than 2,500 gm **
0.29**
-1.22
Source of drinking water
(ref=central/local piped
Well
1.47
0.39
Spring/mineral/rain/snow/river
3.35
1.12
Sanitation facility (
ref=non-standard
Standard
1.18
0.18
Iron pill intake
(less than 90 pills
More than 90 pills
0.68
-0.38
ANC visit
(ref=less than 4 visits
More than 4 visits
0.64
-0.45
LR chi-square
16.71*
115.25***
Pseudo R
2
(Cox and Snell)
0.03
0.21
Note:
***
p≤0.001, **p≤0.01, *p≤0.05