A randomized controlled trial was conducted to evaluate two interventions for prolonging the duration of breast-feeding in a multiethnic sample of 343 low-income urban women. One intervention compared research breast-feeding bedside counseling by a trained counselor, who also made eight telephone calls during the first 3 months of the infant's life, with the routine breast-feeding counseling provided in the hospital by nurses. The other intervention compared commercial discharge packs provided by formula companies with research discharge packs designed to be consistent with the WHO Code of Marketing of Breastmilk Substitutes. When infants were 4 months old, a telephone interviewer unaware of treatment status contacted 95% (324/343) of the women to determine the infants' feeding and health histories. Compared with routine counseling, research counseling delayed the first introduction of solid foods to the infant's diet (P = .03, one-tailed) but did not exert a statistically significant effect on breast-feeding by 4 months' postpartum. Women who received the research discharge pack, compared with those who received the commercial pack, were more likely to prolong exclusive breast-feeding (P = .004, one-tailed), to be partially breast-feeding at 4 months postpartum (P = .04, one-tailed), and to delay the daily use of solid foods in the infant's diet (P = .017, one-tailed). Among the women who received research counseling, the research discharge pack was associated with lower rates of rehospitalization of infants than was the commercial pack (1% v 14%; P = .014, two-tailed). We conclude that in high-risk maternity populations, commercial discharge materials for breast-feeding women should be replaced by materials consistent with the WHO Code.