Facility-based maternal death reviews and criterion-based clinical audit, were introduced in three districts in Malawi in 2006. Can audit and feedback improve the availability, utilisation and quality of emergency obstetric care (EmOC)? Observational study in which emergency obstetric care offered to women who gave birth in 73 health facilities (13 hospitals and 60 health centres) in three districts in Malawi in 2005 (baseline, 41,637 women) was compared to 2006 (43,729 women) and 2007 (51,085 women). The number of comprehensive and basic EmOC facilities did not change over the 3-year period (p for trend=1.000). Although institutional delivery rate decreased in 2006, overall it increased over 3 years (p for trend<0.001) - 31.8% (2005), 31.1% (2006) and 34.7% (2007), and Caesarean section rate was low and did not change (p for trend=0.257) - 1.7% (2005), 1.6% (2006) and 1.5% (2007). There was a significant increase in the met need for EmOC (p for trend<0.001) - 15.2% for 2005, 17.0% for 2006 and 18.8% for 2007. Maternal mortality decreased significantly from 250 per 100,000 women in 2005 to 222 in 2006 and 182 in 2007 (p for trend<0.001). Similarly, the case fatality rate decreased monotonically (p for trend<0.001) - 3.7% (2005), 3.0% (2006) and 1.5% (2007). Audit and feedback can improve availability, utilisation and quality of emergency obstetric care in countries with limited resources. There is need to increase availability of emergency obstetric care by upgrading some health centres to EmOC level through training of staff and provision of equipment and supplies.