6% vs. 23.6%, p < 0.001). The training of staff in BFPCI has been an important strategy for the implementation of this initiative. Caldeira et al., 17 in a controlled study in Minas Gerais, randomized ten units to be trained in BFPCI and ten to constitute the control group, demonstrating the impact of this strategy on the median duration of EBF.

The effects of BFPCI have also been observed on children’s health. Cardoso et al. 18 conducted a study in a basic health unit in Rio de Janeiro comparing pre- and post-certification in BFPCI, and observed an increase in the prevalence of EBF, an increase in routine consultations with asymptomatic infants, and a reduction of consultations whose chief complaint was diarrhea or respiratory infection. Maternal work outside the house was proven to be a risk factor for EBF weaning: it was the variable with the highest intensity of association with the outcome. This result was consistent with that found by CAL 101 Damião19 in Rio de Janeiro, where maternal employment decreased the likelihood of EBF in children younger than 4 months by 41%. Moreover, in a cohort study performed in Pelotas, maternal work increased the likelihood selleck of EBF interruption at 3 months in 76%.20 In 2009, the Family Health

Strategy was beginning to be structured and implemented in the city of Rio de Janeiro, covering only 3.5% of the population, the worst coverage among Brazilian capitals. In 2012, this strategy covered Racecadotril 35% of Rio de Janeiro’s population (approximately 2.2 million people), with massive investments in training of professionals in BFPCI.8 These circumstances may be one of the factors that led the basic health units to generate a prevalence of EBF 10.4% higher than that of the family health units. The assistance provided by the maternity ward staff to teach mothers how to nurse showed no significant association with the outcome. However, a limitation of this study was that this assistance was categorized according to maternal perception; hospital identification was not available, and therefore it was not possible to characterize them as a Baby-Friendly Hospital, a criterion used in other studies to assess the quality of hospital assistance to

promote breastfeeding.21 and 22 In addition to this hypothesis of information bias, another source of explanation for the lack of association observed is that the mothers with more difficulty in the process of establishing breastfeeding in the maternity may have been the target of more encouragement to breastfeed, and these problems with breastfeeding may have continued after the discharge, impairing EBF. Another limitation to be highlighted was the impossibility of controlling for other variables associated with EBF according to the literature, such as maternal characteristics: education, age, primiparity,22 child’s birth weight,23 gender,24 family income,20 and living conditions,25 as in the present study these variables were not collected.