RESUMO
OBJECTIVE: To investigate the identification of maternal deaths at the community level using the reproductive age mortality survey (RAMOS) in all households in which a women of reproductive age (WRA) died and to determine the most concise subset of questions for identifying a pregnancy-related death for further investigation. METHODS: A full RAMOS survey was conducted with the families of 46 deceased WRA who died between 2005 and July 2009 and was compared with the cause of death confirmed by the maternal mortality review committee to establish the number of maternal mortalities. The positive predictive value (PPV) of each RAMOS question for identifying a maternal death was determined. RESULTS: Compared with years of voluntary reporting, active surveillance for maternal deaths doubled their identification. In addition, 4 questions from the full RAMOS have the highest PPV for a maternal death including the question: "Was she pregnant within the last 6weeks?" which had a 100% PPV and a 100% negative predictive value. CONCLUSION: Active identification of maternal mortality at the community level by using a 4-question modified RAMOS that is systematically administered in the local language by health workers can increase understanding of the extent of maternal mortality in rural Ghana.
