The influence of maternal attitudes, perceived social support, and sociodemographic variables on infant-feeding behavior in hospital was examined in 668 black and 511 white primiparae delivering in three metropolitan Washington, D.C., hospitals. The participation rate was 84%. Breast-feeding rates were 84% among white and 49% among black women. The overall breast-feeding attitude measured specific attitudes of "breast is best," "confidence in breast feeding," and "social freedom of breast feeding." Perceived social support from five individuals (baby's father, obstetrician, mother, closest female friend and other relative) was measured. In a stepwise regression model in each ethnic group, the overall attitude explained approximately 35% of the variation in breast or formula feeding behavior in hospital; also, perceived social support and maternal education made significant independent contributions to explaining feeding behavior. Regardless of ethnic group, the "breast is best" attitude and perceived support from the baby's father explained 35% of the variance in feeding behavior after controlling for sociodemographic variables. Both breast and formula feeders had low "confidence in breast feeding" scores and were negatively inclined to believe that "breast feeding allows social freedom." Ethnic differences in attitudes or perceived social support were not found. The obstetrician was perceived as influential and supportive of breast feeding regardless of feeding type; however, in regression analyses perceived social support from the obstetrician was not related to feeding behavior, suggesting a prenatal opportunity for promotion of breast feeding is not being realized.

Among breast feeders 27% of white (n = 431) and 53% of black women (n = 324) used formula supplements in hospital, a factor strongly related to hospital of delivery. Neither maternal attitudes, perceived social support, or sociodemographic variables adequately explained differences between exclusive breast and mixed feeding behavior in hsopital. In a stepwise regression model examining maternal commitment to breast feed, the effect of hospital procedures, and sociodemographic variables, the timing of the first breast feed was strongly associated with formula supplementation. Moreover, breast feeders who delivered by cesarean section (44% n = 755) were significantly more likely to delay the first breast feed. First-time breast feeders are relatively unsure of their breast-feeding ability; thus, hospital procedures should be modified to promote maternal confidence in breast feeding and not undermine the establishment of lactation.