To investigate hospital practices associated with infant feeding during hospital stay and at 1 month of age, all women who were in 17 maternity hospitals on a given day and had delivered a singleton newborn were contacted. Of the 165 eligible women, 153 participated in the survey. Data were collected from the mothers' clinical records and through interview. Odds ratios were calculated to estimate the effect of hospital practices on breastfeeding rates. The main findings are that cesarean section (adjusted odds ratio [OR(adj)] = 3.69; 95% confidence interval [CI], 1.19-11.47), lack of information about the advantages of breastfeeding (OR(adj) = 4.78; 95% CI, 1.50-15.26), and absence or partial absence of rooming-in (OR(adj) = 14.64; 95% CI, 2.24-95.75) increase the risk of complementary breastfeeding during hospital stay. In turn, complementary breastfeeding during hospital stay was the major factor associated with not breastfeeding at 1 month of age (OR(adj) = 6.35; 95% CI, 2.52-16.01). These findings suggest that some hospital practices might have a negative effect on the initiation and continuation of breastfeeding.