Advancing the implementation of quality improvement strategies to reduce maternal mortality using realist evaluation: the case of Maternal Death Reviews in Benin

Project Details


Maternal mortality is highest in sub-Saharan Africa (SSA) with an average maternal mortality rate estimated at 542 maternal deaths per 100 000 live births. Improving the quality of maternity care can significantly reduce maternal and perinatal mortality and morbidity. Although several quality improvement (QI) strategies have been implemented in most SSA countries, these countries are still struggling to improve the quality of care in maternity wards, and maternal mortality has not declined as expected. As part of these strategies, maternal death reviews (MDRs) have been recommended by the World Health Organization since 2013 and were implemented in half of the SSA countries by 2017. MDRs, are defined as qualitative and in-depth examinations of the causes of, and circumstances surrounding, maternal deaths that occur in both health facilities and communities. They hold a key place because they have been shown to have a greater impact on health outcomes than other QI strategies. Recent literature reveals, however, that MDRs have mixed results in lowering maternal mortality and that is due to the differing countries, health districts, and health facilities contexts in which this strategy was implemented. Thus, there is a growing need to understand how MDRs should be implemented to maximize effectiveness in different contexts, especially in low resources settings, either alone, or in combination with other QI strategies.
This Ph.D. proposal aims to address this gap, using the opportunity of a five year European Union-funded project (Action Leveraging Evidence to Reduce perinatal Mortality and Morbidity in sub-Saharan Africa -ALERT). The ALERT project focusses on the quality of maternal and perinatal care and aims to reduce perinatal mortality in 16 hospitals in Benin, Malawi, Uganda, and Tanzania through the co-design and implementation of a multifaceted health system intervention to improve intrapartum care. For this Ph.D. project, I will investigate how, why, and under which circumstances different QI strategies in maternity wards, in limited-resource settings, are effective. Based on these findings, I will more closely examine the case of MDRs to understand which contextual factors affect the effectiveness of MDRs to reduce maternal mortality in low resource settings, alone or integrated into a multifaceted intervention, and how they can be further improved.
I will use the realist evaluation approach as the overall methodological approach. Realist evaluation is a theory-based evaluation approach that seeks to make explicit the root causes and the mechanisms at work in a given intervention. The initial research question is: How do maternity care QI strategies reduce maternal mortality, to which extent, for whom, and under which circumstances? Following the realist evaluation cycle, I will formulate an initial hypothesis, or programme theory, building on a literature/document review, key-informant interviews, and the expertise of the research team. I will then test the initial programme theory using a case study design. I define the case as the design and implementation of MDRs in Benin. I will study this case using mixed methods and collect data through quantitative data extraction from the national databases of maternal deaths and of MDRs, realist interviews with key informants, and non-participant observations. I will first start to collect data in two hospitals (not included in ALERT Project) where MDRs are implemented alone or with other QI strategies but with limited coordinated management. Quantitative data will be analysed with EpiData 15. I will manage qualitative data using NVIVO12 and perform a thematic analysis. I will synthetise the data using the Intervention-Context-Actors-Mechanism-Outcome configuration as an analytical heuristic and identify causal configurations. On this basis, I will refine the initial programme theory. Subsequently, I will perform a second cycle of empirical research. In this second cycle, I will focus on two hospitals where a multifaceted and coordinated intervention including MDRs will be in place (in two ALERT implementing hospitals) using the same methods as discussed above. This second phase will end with a refined programme theory.
Evidence from this study will be used to inform more advanced context-specific QI strategies to reduce in-facility maternal mortality in Benin and SSA. It will generate new evidence on how MDRs are currently operating, either as a single strategy or as a component of a multifaceted and coordinated intervention. This study will also generate methodological learnings, relevant to inform the design, the implementation, and the evaluation of the QI component of the ALERT project.
Effective start/end date1/01/2130/06/22

IWETO expertise domain

  • B680-public-health