In emergency and immigrant health service departments, medical professionals frequently have no language in common with a patient. When no interpreter is available, doctors need another means of collecting patient anamneses. Machine translation was shown to be dangerous and is not available for all languages. BabelDr, a speech-enabled phraselator, was developed for this purpose in a collaboration of the Geneva University Hospitals (HUG) and the Faculty of Translation and Interpreting of Geneva University. In this paper, we focus on the development of the target language resources for the BabelDr system.