Our aim was to report perinatal characteristics of very preterm births before arrival (BBAs) at a hospital, and perinatal and infant mortality rates up to one year, comparing BBAs with births in a hospital.

Materials and Methods
A population‐based cohort study of 22–31 weeks’ gestation births in the state of Victoria, Australia from 1990–2009. BBAs were defined as unintentional births at home or on route to hospital. Perinatal data were obtained from the Department of Health and Human Services, Victoria. Perinatal and infant mortality data comparing BBAs with births in hospitals were analysed by logistic regression, adjusted for gestational age, birthweight and sex.

One hundred and thirty‐three BBAs were recorded: 51 (38%) stillbirths and 82 (62%) livebirths. Compared with births in a hospital, BBAs were less mature (26.3 weeks (SD 2.9) vs 27.7 weeks (SD 2.8), P < 0.001) and a higher proportion were born to teenagers: 13% versus 5% (adjusted odds ratio (aOR) 2.86, P < 0.001). BBAs were significantly more likely to be stillborn (aOR 2.13, 95% confidence interval (CI) 1.41, 3.23, P < 0.001) die within 28 days of livebirth (aOR 2.97, 95% CI 1.54, 5.73, P = 0.001) or die within a year of livebirth (aOR 2.87, 95% CI 1.51, 5.46, P = 0.001) compared with hospital births. Overall, 54 BBAs survived to one year (41% all BBAs, 67% liveborn BBAs), compared with 69% of hospital births (87% of livebirths). Conclusions Very preterm birth before arrival is more common in teenagers and is associated with significantly increased risks of perinatal and infant mortality compared with birth in a hospital.