[Evaluation of the maternal mortality surveillance system in public facilities in Tunisia (1999-2004)].

We report the performance indicators in 2004 of a follow-up on the system for recording maternal deaths which was established in 1999. The system was operating in 69.8% of public hospitals, and 96% of maternal deaths investigations were completed. In 69.8% of maternal deaths there was a direct obstetric cause. Haemorrhage was the major cause of maternal death (30.8%), followed by eclampsia (11%). The proportion of avoidable (certain or possible) deaths was 75.3%. There were problems in evaluation of risk presented by women and inadequate follow-up during the postpartum period and delay in appropriate treatment. Incomplete documentation and difficulty in ascertaining avoidability were problems faced by the regional follow-up committee.