Healthcare interpreting, performed via tele/video-conference or face-to-face interactions is complex. Research in healthcare interpreting has contributed to our understanding of this practice (Metzger 1999; Davidson 2001; Angelelli 2004, 2011, 2012; Baraldi/Gavioli 2012; Meyer 2012). Access to cross-cultural/linguistic interactions between provider/patient mediated by interpreters is essential to study intercultural/linguistic healthcare communication. Access to naturalistic data, however, is not always feasible. Therefore, researchers rely more and more on secondary data for analysis. This paper discusses ethical, practical and scientific dilemmas experienced when assessing the feasibility of turning ethnographic data into data for corpus studies. Firstly, after an introduction and a concise review of the principles underlying ethnography, the original studies are explained briefly to contextualize the data. These studies are: a) an ethnography (Spanish-English) of a medical interpreting unit and b) two case studies (Cantonese/Hmong - English) conducted in a total of three public hospitals in the United States. Secondly, a discussion on using data for a different purpose than the original one, and the resulting ethical, practical and scientific dilemmas will be presented. The goal is to reflect on and examine if the opportunities to advance science may outweigh the issues raised in this paper and if it would be ethical to proceed.