Objective. To analyze changes in postpartum hemorrhage over a 10-year period from 1998 to 2007, and to explore factors associated with severe hemorrhage. Design. Retrospective cohort study, prospectively collected information. Setting. Stavanger University Hospital, a secondary referral center, Norway. Population. An unselected population of 41,365 women giving birth at the hospital. Methods. We analyzed changes over time in mean postpartum hemorrhage, severe postpartum hemorrhage and associated factors. Estimated blood loss >1,000 ml was defined as severe hemorrhage. Data were collected from the hospital's database. Main outcome measures. Severe postpartum hemorrhage and obstetric interventions. Results. We observed an increase in severe hemorrhage during the study period. After cesarean sections, the risk of severe hemorrhage was twice the risk of severe hemorrhage after vaginal deliveries (5.9%; 95% CI 5.3–6.6 vs. 2.8%; 95% CI 2.6–2.9). The most important factors associated with severe hemorrhage following vaginal deliveries were twin deliveries (OR 6.8), retained placenta (OR 3.9) and inductions of labor (OR 2.2). For cesarean sections, twin deliveries had the strongest association with severe hemorrhage (OR 3.7) followed by general anesthesia (OR 3.0). Obstetric interventions became more frequent; elective cesarean sections increased from 2.4 to 4.9%, acute cesarean sections from 5.5 to 8.9%, operative vaginal deliveries from 9.3 to 12.5%, inductions of labor from 14.3 to 15.8% and augmentations of labor from 5.8 to 29.3%. Conclusions. The incidence of severe postpartum hemorrhage increased, and this may be related to more frequent use of obstetric interventions.