Factors associated with the duration of breastfeeding: analysis of the primary and secondary responders to a self-completed questionnaire.

A cohort of 1192 consecutive newborn infants was followed prospectively for factors possibly affecting the length of time they were breastfed. Following the procedure of a double-blind test, one-third of the cohort received Credé prophylaxis at age 2 h. The duration of breastfeeding (sole or partial) was recorded up to age 6 months and there was a 100% follow-up. Multivariate proportional hazards regression analysis (Cox) of the whole cohort showed that babies being delivered between 21.00 and 24.00 h were associated with a shorter duration of breastfeeding (rate ratio = 1.37, 99% confidence interval = 1.05-1.78). Mother's age (under 21 years), marital status (unmarried) and birthweight (inversely) were factors also independently associated with shorter breastfeeding duration. Boys were breastfed for a shorter time than girls (p < 0.05). In univariate analyses only, the first-born babies had a significantly shorter breastfeeding time, and purulent eyes in the first 24 h was a factor of borderline significance (p < 0.05). Educational level, socioeconomic status and smoking habits of the mothers were not investigated in this study. Owing to the lack of regulations in place at the time of the study (1981-82), it was possible to differentiate between the mothers who responded spontaneously to the self-completion questionnaire (primary responders, 68.5%) and those who required one or two reminders. Short breastfeeding time was the strongest predictor of being a secondary responder, followed by being very young or unmarried. Approaching the secondary responders reduced the prevalence of breastfeeding at 6 months by 5% (from 53.8% to 48.8%).