What you count is what you target: the implications of maternal death classification for tracking progress towards reducing maternal mortality in developing countries.

The first target of the fifth United Nations Millennium Development Goal is to reduce maternal mortality by 75% between 1990 and 2015. This target is critically off track. Despite difficulties inherent in measuring maternal mortality, interventions aimed at reducing it must be monitored and evaluated to determine the most effective strategies in different contexts. In some contexts, the direct causes of maternal death, such as haemorrhage and sepsis, predominate and can be tackled effectively through providing access to skilled birth attendance and emergency obstetric care. In others, indirect causes of maternal death, such as HIV/AIDS and malaria, make a significant contribution and require alternative interventions. Methods of planning and evaluating maternal health interventions that do not differentiate between direct and indirect maternal deaths may lead to unrealistic expectations of effectiveness or mask progress in tackling specific causes. Furthermore, the need for additional or alternative interventions to tackle the causes of indirect maternal death may not be recognized if all-cause maternal death is used as the sole outcome indicator. This article illustrates the importance of differentiating between direct and indirect maternal deaths by analysing historical data from England and Wales and contemporary data from Ghana, Rwanda and South Africa. The principal aim of the paper is to highlight the need to differentiate deaths in this way when evaluating maternal mortality, particularly when judging progress towards the fifth Millennium Development Goal. It is recommended that the potential effect of maternity services failing to take indirect maternal deaths into account should be modelled.