Too many women in low- and middle-income countries (LMICs) still die as a direct result of pregnancy and childbirth. We know from experiences in high-income countries that the vast majority of these deaths are preventable. Most experts believe that strengthening various aspects of the health system (as conceptualized in the six WHO building blocks, which include the health workforce; health information systems; leadership and governance; and actual service delivery) is the solution.

It is still, however, unclear which building blocks are most important, which interventions within the building blocks are best value for money, and how interventions in different building blocks influence one another. Another aspect that is critical for improving maternal health but is insufficiently emphasised in the building blocks is enabling patients to demand good access to highquality services.

We postulate that the key steps needed to improve services in pregnancy include holding maternal health managers accountable for these services and encouraging community participation to thus increase patient demand. Moreover, much more is needed to shore-up the skills of public-sector managers and how they function in teams that work to improve these services. Finally, we hold that services would be improved through better use of information routinely collected by maternal health services, especially learning from instances of maternal death.

To test this hypothesis, we will thoroughly review the available knowledge on these topics in the academic literature (for all low- and middle-income countries) and in the grey literature (for South Africa, Rwanda and three additional, carefully selected African countries). Then, in Rwanda and South Africa, we will document illustrative cases of successes and failures in maternal health services, to identify opportunities to improve the way these services are provided, and to enhance workers‟ motivation and leadership.

We will study services for treatment of HIV disease in pregnant women and for providing care during emergencies in pregnancy or childbirth. Through working closely with policy leaders and civil society, throughout the project, we will take joint actions to improve maternal services. Practically, this means that teams of policy makers and researchers in Rwanda and South Africa will apply knowledge learnt in the project and attempt to markedly improve the way maternal services are organised. In both countries, the major reforms underway and the relative inclusiveness of policy processes, mean that presently there are major opportunities to positively influence policy directions and their implementation.