Using the new ICD-MM classification system for attribution of cause of maternal death—a pilot study CA Ameh, a A Adegoke, a RC Pattinson, b N van den Broek a a Centre for Maternal and Newborn Health, Liverpool School of Tropical Medicine, Liverpool, UK b SA MRC Maternal and Infant Health Care Strategies Unit, Department of Obstetrics and Gynaecology, University of Pretoria, Pretoria, South Africa Correspondence: CA Ameh, Centre for Maternal and Newborn Health, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK. Email Charles.Ameh@lstmed.ac.uk Accepted 6 June 2014. Understanding the causes of and factors contributing to maternal deaths is critically important for development of interventions that reduce the global burden of maternal mortality and morbidity. The International Classification of Diseases—Maternal Mortality (ICD-MM) classification of cause of death during pregnancy, childbirth and the puerperium was applied to data obtained from maternal death reviews (MDR) for 4558 maternal deaths from five countries in sub-Saharan Africa. None of the data sets identified type of maternal death. Information obtained via MDR is generally sufficient to agree on classification of cause of death to the levels of type and group. The terms ‘underlying cause of death’ and ‘contributing conditions’ were used differently in different settings and a specific underlying cause of death was frequently not recorded. Application of ICD-MM resulted in the reclassification of 3.1% (9/285) of cases to the group ‘unanticipated complications of management’, previously recorded as obstetric haemorrhage or unknown. An increased number of cases were assigned to the groups pregnancy-related infection (5.6–10.2%) and pregnancies with abortive outcome (3.4–4.9%) when a clear distinction was made between women who died ‘with’ HIV/AIDS of obstetric causes (direct maternal death) and AIDS-related indirect maternal deaths (group ‘non-obstetric complications’). Similarly, anaemia and obstructed labour were more frequently identified as contributing factors than underlying cause of death. It would be helpful if MDR forms could have explicitly stated variables called: type, group and underlying cause of death as well as a dedicated section to the most frequently occurring contributing conditions recognised in that setting. Keywords ICD-MM, maternal death classification, maternal mortality. Please cite this paper as: Ameh C, Adegoke A, Pattinson RC, van den Broek N. Using the new ICD-MM classification system for attribution of cause of maternal death—a pilot study. BJOG 2014; 121 (Suppl. 4): 32–40. Introduction From 1996, 5-yearly estimates of maternal mortality have been provided; in 2005, just under 536 000 maternal deaths were estimated to have occurred globally, in 2008 around 343 000 women died and the most recent estimate amounts to 289 000 deaths for 2013. 1–3 Overall therefore, there is progress but there is concern that this is not fast enough to meet the Millennium Devel- opment Goal (MDG) 5a target of a 75% reduction in maternal mortality by 2015 compared with the 1990 level. 4,5 In addition, it must be noted that for many coun- tries accurate data are not available and we continue to rely on estimates. Countries where maternal mortality is high are often those with the least accurate data. There is also progress with regard to understanding why women die during pregnancy, childbirth and the postnatal period. An early approach to analysis of this was the ‘three-delay framework’, which is still applied in many settings to identify why women experience delays when they need health care, at which level these delays occur, and the contributing factors. 6 Whereas information regard- ing delay in access to care is useful to analyse how access can be improved, analysis of the medical causes of deaths is also needed to be able to design and implement effective healthcare packages. 32 ª 2014 Royal College of Obstetricians and Gynaecologists DOI: 10.1111/1471-0528.12987 www.bjog.org Review article