Early detection of maternal deaths in Senegal through household-based death notification integrating verbal and social autopsy: a community-level case study

Background: Reliable detection of maternal deaths is an essential prerequisite for successful diagnosis of barriers to care and formulation of relevant targeted interventions. In a community-level case study, the use of household-level surveillance in Senegal unveiled an apparent increase in maternal deaths, which triggered a rapid-cycle collaborative response to implement a multipronged set of quick-win and sustained interventions intended to improve quality care. Methods: Part of a multi-country effort, the Millennium Villages Project is implementing a routine community-level information system in Senegal, able to detect maternal deaths in real-time and uncover clinical and social factors contributing to mortality. Within this geographically demarcated area of approximately 32 000 inhabitants, with a well-structured health system with patient referral services, deaths were registered and notified by community health workers, followed by timely verbal and social autopsies. Using the Pathway to Survival conceptual framework, case analysis and mortality reviews were conducted for evaluation and quality improvement purposes. Results: The estimated maternal mortality rates rose from 67/100000 births in 2009 (1 death), to 202/100000 births in 2010 (3 deaths) and 392/100000 births (5 deaths) in 2011. Although absolute numbers of maternal deaths remained too small for robust statistical analysis, following verbal autopsy analyses in 2011, it became evident that an unexpectedly high proportion of maternal deaths were occurring at the referral hospital, mostly post-Caesarian section. Inadequate case management of post-partum haemorrhage at the referral hospital was the most frequently identified probable cause of death. A joint task team systematically identified several layers of inefficiencies, with a potential negative impact on a larger catchment area than the study community. Conclusions: In this study, routine community-based surveillance identified inefficiencies at a tertiary level of care. Community-level surveillance systems that include pregnancy, birth and death tracking through household visits by community health workers, combined with verbal and social autopsy can identify barriers within the continuum of maternal care. Use of mHealth data collection tools sensitive enough to detect small changes in community-level mortality trends in real-time, can facilitate rapid-cycle quality improvement interventions, particularly when associated with social accountability structures of mortality reviews