Objective: To identify the population risk factors in emergency situations, we studied breastfeeding status as a predictor for child mortality during a war in Guinea-Bissau.
Methods: Data on breastfeeding status are routinely collected by the surveillance system of the Bandim Health Project in Bissau. We used data collected during a 3-month period prior to the war in Guinea-Bissau in June 1998 to assess the impact of breastfeeding status on mortality in an emergency. We compared the war cohort with two cohorts of children who had had their breastfeeding status assessed in a similar way by the surveillance system in the 3 months prior to June 1996 and June 1997. As very few are weaned prior to 9 months of age and the median age of weaning is 22 months, we assessed the risk of dying over a 3-month period for breastfed and weaned children aged 9-20 months.
Results: Controlling for age, weaned children experienced a sixfold higher mortality [mortality rate (MR) = 5.73 (95% CI 2.40-13.71)] during the first 3 months of the war compared with children still breastfeeding. In the two control cohorts from 1996 and 1997, weaned children did not have higher mortality than the breastfed children over a similar 3-month period. Mortality in weaned children was five times higher [MR = 4.96 (1.44-16.63)] during the first 3 months of the conflict than in a similar group of weaned children from early June 1996 and June 1997, whereas there was no significant difference in mortality between breastfed children during the conflict and the preceding years [MR = 1.46 (0.84-2.55)]. Control for other background factors, including living with mother, gender, ethnic group, mother's schooling and district, did not change these differences.
Conclusion: The protective effect of breastfeeding against infections may be particularly important in emergencies. Continuing or recommencing breastfeeding should be emphasized in emergency situations.