The common comparison of depression to diabetes enables the construction of depression as a nonstigmatizing chronic illness that requires medication. We explore, through the use of discourse analysis, how both long-term users of antidepressants and family physicians invoked this analogy in research interviews. Specifically, we show how these participants explicitly or implicitly challenged the aptness of the depression-diabetes analogy as framed either within a generic (and presumably type 1) conception of diabetes or within the model of type 2 diabetes. These challenges include demonstrating how the elements or inferences of the analogy do not correspond, and how the analogy does not have its intended effects. We consider the implications of the unraveling of this analogy for the construction of depression as a chronic medical condition, for the supposed ease of prescribing and taking antidepressants, and for the reduction of stigma.