RECENT EVOLUTION OF INFANT-MORTALITY AT B AMAKO (MALI)

A follow-up study of 4, 575 live births to 4,718 pregnant women in the Bankoni neighbourhood of Bamako is the basis for estimating the vario us risk factors of infant mortality. The cohort was formed in 1989-199 2, and the last birth was in August 1993. Neonatal, post-neonatal and infant death rates were estimated at, respectively, 15, 40 and 55.6 pe r 1,000 live births. A survival analysis was used to identify a number of independent predictors of infant mortality : twin births (RR = 5,0 [3.3-7.7]), primiparity (RR = 1.5 [1.1-2.1]), length of residence in a neighbourhood less than 5 years (RR = 1,4 [1.1-1.9]), children born between 1991 and 1993 (RR = 1.4 [1.1-1.9]) and during the hot, rainy s eason (RR = 1.4 [1.1-1.9]). A selection bias caused by the proportion of children born live but not seen again at their birthday is examined . This estimate of infant mortality, based on a short and recent perio d, is lower than earlier estimates from the 1980s for Bamako city. But it is comparable to urban infant mortality in the period 1983-1992 fo r Senegal, where a strong fall has been observed since the mid 1980s. A recent estimate for the third-largest city in Mall (Sikasso) suggest s the same pattern as Bamako.