OBJECTIVE: To evaluate whether providing doulas during hospital-based labor affects mode of delivery, epidural use, breast-feeding, and postpartum perceptions of the birth, self-esteem, and depression. METHODS: This was a randomized study of nullipara enrollees in a group-model health maintenance organization who delivered in one of three health maintenance organization-managed hospitals; 149 had doulas, and 165 had usual care. Study data were obtained from the mothers’ medical charts, study intake forms, and phone interviews conducted 4-6 weeks postpartum. RESULTS: Women who had doulas had significantly less epidural use (54.4% versus 66.1%, P < .05) than women in the usual-care group. They also were significantly (P < .05) more likely to rate the birth experience as good (82.5% versus 67.4%), to feel they coped very well with labor (46.8% versus 28.3%), and to feel labor had a very positive effect on their feelings as women (58.0% versus 43.7%) and perception of their bodies' strength and performance (58.0% versus 41.0%). The two groups did not differ significantly in rates of cesarean, vaginal, forceps, or vacuum delivery, oxytocin administration; or breast-feeding, nor did they differ on the postpartum depression or self-esteem measures. CONCLUSION: For this population and setting, labor support from doulas had a desirable effect on epidural use and women's perceptions of birth, but did not alter need for operative deliveries.