Anthropological and STS scholars of biomedical work have traditionally explored contexts where inconsistencies and lacunas in diagnostic knowledge-production are problematic for medical practitioners, and such scholars have consequently focused on the social and political processes by which such epistemic uncertainties are resolved. This article draws on ethnographic material from a Papua New Guinean hospital where diagnostic uncertainty is not rendered problematic and where the open-endedness of the diagnostic process gives rise to new forms of medical expertise and practice. The paper focuses on the medical record as an artefact of not-knowing that both documents and performs uncertainty as a valuable resource. It shows that medical records can operate as either technologies of 'opening' that multiply opportunities for pragmatic action within a hospital space or as technologies of 'closure' that move people and documents between spaces. Practices of not-knowing and knowing are therefore shown to be interdependent and interchangeable 'moments' of bureaucratic-biomedical work. © SAGE Publications 2011.