Methods We use nationally representative data from eight countries in Latin America and the Caribbean to document changes in breastfeeding duration between 1986 and 2005, and separate the overall change into the portion attributable to changing population characteristics and the portion resulting from changing breastfeeding behaviour within population subgroups.
Results Breastfeeding duration increased in six out of the eight countries and the changes observed are largely explained by changing behaviour within population subgroups rather than changing population characteristics. Changes in breastfeeding duration did not tend to be equitably distributed, but in four countries (Bolivia, Brazil, Colombia and Peru) the population subgroups whose children are most at risk for mortality and increased morbidity from not being breastfed were least likely to show improvements in breastfeeding duration. Between 1986 and 2004 in Peru, breastfeeding duration declined by 0.6 months among rural women while increasing by 9.7 months among urban women; it increased by 6.3 months among women with prenatal care but only by 3.7 months among women with no prenatal care. Changes in breastfeeding in Guatemala and Haiti tended to favour the well-off compared with the poor, though not consistently. In Nicaragua changes in breastfeeding duration tended to favour the less well-off.
Discussion While promoting breastfeeding is a must for all women, to maximize its benefits for child survival and health, additional efforts are needed to reach poorly educated and rural women with little access to health care.