Objective: To assess the frequency and outcomes of preterm hospital admissions during pregnancy, with a focus on transfers to higher levels of care. Design: Population-based cohort study using linked population data. Setting and subjects: Women who were admitted to hospital in weeks 20-36 of pregnancy (preterm) and gave birth to a liveborn singleton infant in New South Wales during 2001-2008. Main outcome measure: Numbers of preterm admissions of pregnant women who were discharged without giving birth, were transferred to higher care, or who gave birth. Results: 110439 pregnancies (16.0%) involved at least one preterm admission. After their initial preterm admission, 71.9% of women were discharged, 6.3% were transferred and 21.8% gave birth. Median gestational age at admission was 33 weeks and median time to discharge, transfer or giving birth was 1 day. Most women who were transferred or who gave birth had been admitted for preterm rupture of membranes or preterm labour. Of the women who were admitted or were transferred with suspected preterm labour, only 29% and 38%, respectively, gave birth. Compared with other admitted women, women having a first birth, public patients and those living in areas of low socioeconomic status were more likely to be transferred or to give birth. As gestational age increased, the proportion of women transferred decreased and the proportion giving birth increased. Infants born after maternal transfer had lower gestational age and more adverse outcomes than those born without maternal transfer. Conclusions: Preterm hospital admission affects one in six women with singleton pregnancies. Methods that could improve assessment of labour status have a large potential to reduce the burden on maternity services. The increased morbidity for infants born after maternal transfer suggests women with highrisk pregnancies are being appropriately identified.