In many low- and middle-income countries, where an increasing percentage of women are giving birth in facilities, poor healthcare worker practice is seen as one of the key obstacles to providing quality maternal and newborn health care services. In this PhD by Prior Publication, I use four illustrative papers, linked by a 15,000-word text, to argue that to understand and improve healthcare worker practice it is important to recognise that healthcare workers, like all of us, are not “robots” or “angels” but complex human beings, embedded in complex hierarchical health systems and wider communities, with a range of identities and motivations that shape their everyday practice. Paper One, an ethnography of maternity wards in South Africa, explores how policy to improve public finance management was implemented in a way that replicated existing hierarchies, alienated staff, and undermined improvements in quality of care. Paper Two, on work in South Africa, Russia, Uganda, and Bangladesh, highlights the importance of recognising that what is officially expected to happen in health systems often differs from reality, and that this ambiguity lies in the lived reality of many healthcare workers and contributes to implementation failure. Paper Three, a mixed methods process evaluation of a social franchise intervention in India, used ethnographic methodologies to show how the logic and motivation of private healthcare providers distorted the implementation process. Paper Four documents experience with primary health care clinic managers in Senegal using reflexive diaries and how their engagement with the research process was initially shaped by the health system context where they felt constantly audited and “reported up” the system. A continuous theme in all four papers, and my research during the past 20 years in general, is that researchers who are working to better understand healthcare worker practice and how to improve it must take time to understand the health system and wider context in which healthcare workers operate to be able to ask the right questions. The health system context—often bureaucratic, top down, punitive, and not inclusive of healthcare workers’ voices—also shapes how healthcare workers react and respond to researchers and their questions, and this must be taken account of at all stages of the research process.