In this paper, we present an ongoing project that has used video ethnographic methods to investigate the intersections of safety, communication and space in clinical contexts. This research has been carried out in intensive care units, and in general medical and general assessment wards. The approach has involved interviews and observations to narrow down the foci for filming, followed by reflexive sessions where participants and researchers viewed clips of recorded footage to make sense of it together. Unlike models of visual research which separate researchers and participants in the processes of producing and interpreting visual data, this project explicitly set out to engage both researchers and participants dialogically and iteratively in the creation and interpretation of visual data. Practically, this co-production has required ongoing reflexive attention to shifting power relationships between researchers and participants and among participants. Furthermore, methodological concerns such as rigor, validity, ethics and feasibility are necessarily considered, assessed and determined primarily during fieldwork and data creation, rather than preemptively. Viewing video data reflexively provided opportunities to critically appraise practices, examine practices that were previously taken-for-granted and provide a forum in which clinicians’ views could be voiced and contested. We describe examples of these research practices, and discuss their limitations, risks, and also their powerful potentialities for enabling change and new insights for both researchers and participants in the contexts that we study.