The care of pregnant women and neonates in peripheral hospitals in many developing countries is in a critical state. Through a retrospective analysis we assessed the effects of the introduction of standardised protocols in obstetric and neonatal care (implementation from 1998 onwards) on perinatal and neonatal outcomes of all deliveries over seven years (1996–2002) at a first-referral hospital in rural Tanzania. In all, there were 18026 deliveries (18316 live births and 606 stillbirths). Perinatal mortality rates (PMR) varied from 42.8–54.5/1000 live births during the years. Early neonatal mortality rates (eNMR) fell from 21.9/1000 live births in 1996 to 14.8/1000 live births in 2002 (all p>0.05). Fresh stillbirth rates decreased over time (p=0.041), however macerated stillbirth rates increased during the second half of the period (p=0.067). Sixty-two to seventy-two percent of eNMR occurred on the first day of life (p<0.001). Maternal mortality ratio declined from 729/100000 live births in 1996 to 119/100000 live births in 2002 (p=0.002). Our clinical project was associated with a reduction of PMR and eNMR (and maternal mortality ratios), but with considerable fluctuations during the years. Improving obstetric and neonatal care in the hospital setting in developing countries is essential, but needs long-term commitment and support.