Objectives: To investigate why physicians label end-of-life acts as either 'euthanasia/ending of life' or 'alleviation of symptoms/palliative or terminal sedation', and to study the association of such labelling with intended reporting of these acts. Methods: Questionnaires were sent to a random, stratified sample of 2100 Dutch physicians (response: 55%). They were asked to label six hypothetical end-of-life cases: three 'standard' cases and three cases randomly selected (out of 47), that varied according to (I) type of medication, (2) physician's intention, (3) type of patient request, (4) patient's life expectancy and (5) time until death. We identified the extent to which characteristics of cases are associated with physician's labelling, with multilevel multivariable logistic regression. Results: The characteristics that contributed most to labelling cases as 'euthanasia/ending of life' were the administration of muscle relaxants (99% of these cases were labelled as 'euthanasia/ending of life') or disproportional morphine (63% of these cases were labelled accordingly). Other important factors were an intention to hasten death (54%) and a life expectancy of several months (46%). Physicians were much more willing to report cases labelled as 'euthanasia' (87%) or 'ending of life' (56%) than other cases. Conclusions: Similar cases are not uniformly labelled. However, a physicians' label is strongly associated with their willingness to report their acts. Differences in how physicians label similar acts impede complete societal control. Further education and debate could enhance the level of agreement about what is physician-assisted dying, and thus should be reported, and what not.