If the term 'evidence-based medicine' conveys more than is conveyed by the word medicine, then there must be a way to distinguish between evidence-based medicine and non-evidence-based medicine. In particular, there must be a logically acceptable way to classify medical decisions as justified or unjustified by scientific evidence. In this essay I examine the nature of medical theories, the nature of the evidence that is produced by empirical tests of medical theories, and the relation of medical decisions to both. I conclude that attempts to classify medical decisions as justified or unjustified by scientific evidence have no foundation in logic and that the term 'evidence-based medicine' is logically indistinguishable from the term 'medicine'. The use of the term 'evidence-based medicine' calls for a new type of authoritarianism in medical practice.