OBJECTIVE: To assess the feasibility of establishing a maternal mortality audit in Lebanon. DESIGN: A facility-based reproductive age mortality study (RAMOS). METHODS: Records of reproductive-age female deaths over an eight-year period in three hospitals in Beirut, Lebanon, were reviewed. MAIN OUTCOME MEASURES: Quality of data sources, comprehensiveness of medical charts and accessibility of information. RESULTS: Review of records was feasible. Cross-checking three data sources identified missed cases and coding and data entry errors. The quality and accessibility of data varied between hospitals. The maternal mortality rate was 39/100,000 live births and there were 55 pregnancy-related deaths/100,000 live births. Lack of antenatal care, delay in seeking care and over-intervention on the part of the medical team were among problems identified. CONCLUSIONS: A facility-based approach is a potential tool for conducting a national maternal mortality audit in a developing country like Lebanon. Computerized medical records and mandatory participation of hospitals are prerequisites for success. This would require the government to develop a cohesive national policy on reducing maternal mortality.