60 / 89
МОНГОЛЫН ХҮН АМЫН СЭТГҮҮЛ Дугаар (367) 20, 2011
59
However, when looked at the odds ratio of monthly
income per person, the odds of having child
mortality was 52 per cent lower for those who had
more than 102,535
₮
month than those who had less
than 102,535
₮
. As for child mortality by mother’s
education, those who had complete secondary
and tertiary education experienced child mortality
1.1 times lower than those who had incomplete
secondary education did.
For the second model, all other independent
variables were added into the model. This time
Pseudo R
2
increased from 0.03 in the previous
model to 0.21, as well as LR chi-square increased
dramatically from 16.71 in the first model to
115.25 in the second model. The result of odds
ratio logistic regression was that a child whose
mother’s age was 30 years old or more was 81
per cent less likely to experience child mortality
than the child whose mother’s age was less
than 20 years old at p≤0.01 level. None of place
of residence as regions, monthly income per
person and mother’s education had a statistically
significant impact on child mortality. However,
in the model two, the size of the odds of mothers
experiencing child mortality in Western province
decreased by two times from that of in the previous
model.
Breastfeeding had the most statistically significant
impact on child survival. Odds of child mortality
for those who ever breastfed were 96 per cent
lower than that of those who never breastfed. Birth
order did not statistically significantly associate
with child mortality. Birth order of the child
did not have statistically significant impact on
child mortality. Compared to those children who
weighed less than 2,500 gm, odds of not dying
for those who weighed more than 2,500 gm were
71 per cent lower. Gender of child was statistically
significant to child mortality; a female child was 48
per cent less likely to die than a male child.
Neither source of drinking water nor sanitation
facility had a statistically significant relationship
with child mortality. Compared to those who got
their water from central and local pipe, those who
got from well had odds of having child mortality
were 1.47 times higher, while those whose drinking
water source from spring/mineral/rain/snow/river
experienced child mortality 3.35 times higher
than the former. The odds of child mortality for a
mother without a standard sanitation facility were
about 1.2 times higher than that of a mother with
standard sanitation.
Health care service mothers received during their
pregnancy did no have statistically significant
impact on child mortality. However, the odds of
child mortality for those mothers who had more
than four antenatal care visits were 36 per cent
lower than for those mothers who had less than
four antenatal care visits. Mothers who took more
than 90 iron pills a day were 32 per cent less likely
to experience child mortality than those who took
less than 90 iron pills a day.
DISCUSSION
There is a big difference in terms of socio-
economic disparity and availability of health
facility between Ulaanbaatar the capital city and
other regions that are considered as rural areas
(MoH, 2005; Hill, Dodd et al., 2006). Ironically,
place of residence had neither statistically
significant relationship with nor impact on
child mortality. It was found that cases of child
mortality happened to the mothers whose age
were less than 20 years old; and this finding was
consistent with experts’ observation about J shape
relationship between child mortality and mother’s
age (Bongaarts, 1987; Davis, 1988; Sullivian
et al, 1994). Only a quarter of the mothers had
monthly income per capita more than 102,535
₮
;
however, the income did not have any significant
impact on child mortality which was different
than the views about the fact that income had
direct impact on child mortality (Kabubo et
al., 2006; Mahgoub et al., 2006). Even though
there was no statistically significant impact of
mother’s education on child mortality, the inverse
relationship between mother’s education and
child mortality was observed in the result which
was common finding the experts (Caldwell, 1979;
Cochrane, 1980). Birth order 1 and 4 and higher
had significant relationship with child mortality
which was supported by other studies (Ahmed,
1986; Bongaarts, 1987). However, later on when
the variable put in the model, birth order turned
out not to have statistically significant impact on
child mortality. Similarly source of drinking water
had significant relationship with child mortality,
especially those mothers whose source of drinking
water was from mineral/spring/river/snow/rain
experienced more child mortality because this