M. Deeb et al., CHILD SURVIVAL IN BEIRUT DURING WARTIME - TIME TRENDS AND SOCIORELIGIOUS DIFFERENTIALS, International journal of epidemiology, 26(1), 1997, pp. 110-119
Background. Child mortality estimates in Beirut are presented for the
late 1970s and the mid 1980s, and changes in socio-religious different
ials of mortality across time are investigated. Methods. Baseline info
rmation was obtained from maternity registries in Beirut in 1984 and 1
991. Age of mother, number of children ever born and number of childre
n alive, hospitalization class (1, 2 or 3 within each hospital, depend
ing on the room rate, the services and the doctor's fees), and religio
n of newborn was recorded. Brass and Macrae's technique was used to co
nvert the proportion dead among children ever born to mothers in age g
roup 30-34 into (5)q(0) estimates for the reference periods 1978 (1984
data) and 1985 (1991 data). Using hospitalization class as a proxy fo
r social class, religious differentials in child mortality were explor
ed by logistic regression analysis within each period. Results. Overal
l, (5)q(0) is estimated at 46 per 1000 in the late 1970s, and at 36 pe
r 1000 in the mid 1980s. The religious differentials in (5)q(0) mortal
ity estimates were quite large in the former period, with a risk for M
uslims relative to Christians of 1.53, which declined to 1.35 in the l
atter period. The regression analysis of the proportion dead among chi
ldren ever born; (1) confirmed the magnitude of the religious differen
tials in child mortality, and their reduction over time; (2) demonstra
ted the existence of an interaction between religion and social class,
as the religious differentials in child mortality were found to be hi
ghly prevalent in the lower social classes, and almost insignificant i
n the middle and upper ones; (3) highlighted the decisive role of diff
erential fertility in generating differential chili mortality, with hi
gher fertility and higher child mortality in Muslims, and with the dis
appearance of religious differentials in child mortality after adjustm
ent for fertility. Conclusion. The findings confirm the hypothesis of
a reduction over time of religious-based child mortality differences i
n the capital city of Lebanon, in agreement with the pattern found for
religious-based fertility differences. Comparisons with other countri
es in the region indicate that the war in Lebanon has slowed down the
decline in child mortality, causing the country to loose its priviledg
ed position among Arab countries.