Abstract

PIP: Studies of the prevalence and duration of breastfeeding as a function of possible determinants in society, the family, or factors related to the mother or the infant are studied by multivariate approaches that allow for simultaneous interpretation of associations and interactions between the variables. Examples of 3 such approaches are given using longitudinal data taken from studies of infants in 3 distinct areas of Sweden between 1979-81. The educational level and smoking habits of the mother were shown to be predictors of the prevalence of breastfeeding in children 3 months of age. The educational background of the mother, her smoking habits, age, and whether her child had been admitted to a pediatric ward during the neonatal period were also found to be significantly associated with breastfeeding patterns. (author's)