A review of methodology and tools for measuring maternal mortality in humanitarian settings.

Estimation of maternal mortality ratio (MMR) in humanitarian settings ('settings of conflict, displacement and natural disaster') is challenging, particularly where communities have dissolved and geographical areas are inaccessible. During humanitarian events, the reproduction of maternal mortality figures by the media is common, and are often based on inaccurate reports. In light of such uncertainties and challenges, the aim of this article was to review and appraise the methodology and data collection tools used to measure MMR in humanitarian settings. A critical review of both grey and peer-review publications was conducted, focussing on articles published from January 1995 until December 2016. In the final review, articles that provided an estimate of MMR from a humanitarian setting were included. The assessment of study quality was based on an adapted framework for the quality of mortality studies in humanitarian settings. Overall, 13 peer-review publications and one grey publication were included in the final review. These were grouped according to settings: camp, clinic, household and census. Studies varied in their definition of MMR, and few studies objectively defined the humanitarian setting. Household-based studies were based on retrospective designs and on the recall of surviving family members. Although many studies attempted to purposively sample the populations afflicted, there was substantial evidence of selection bias; few studies were able to confirm the maternal deaths through medical certificates, or attempted to visit homes to re-inquire about deaths using verbal autopsy. The variation in methods and tools applied suggest that maternal mortality estimates are more likely to be markedly different from the true unknown level. The implications are that a standardized methodology and tools are necessary: that are consistent in definition, use a representative sample where possible, attempt to triangulate and validate data sources, and reconfirm deaths through household visits with informant interviews.