Research issues in the study of difficult-to-treat depression

Recent clinical investigations have revealed that a substantial proportion of patients with depression do not have a satisfactory therapeutic outcome with an initial treatment attempt, or even with subsequent attempts. The preferred outcome is complete symptom remission. In some cases one must accept a clinically significant symptom reduction that is short of full symptom remission. Depression may be difficult to treat because of the nature of the condition itself; factors that interfere with the proper delivery of optimal treatment (such as poor adherence or underdosing); associated concurrent Axis I, II, or III disorders; or the lack of effective treatments. If treatment is optimally delivered and an unsatisfactory outcome occurs, treatment resistance is said to be present. This article reviews critical elements to consider when designing controlled trials of treatments for treatment-resistant major depressive disorder. Such elements include the definition of treatment resistance, methods to document previous failed treatment trials, selection of appropriate research populations, measurement of relevant clinical outcomes, tactical issues in delivering the experimental treatment (such as dosages and durations), and trial design choices (for example, switching vs. augmentation studies). Careful consideration of these issues should improve the interpretability and generalizability of findings obtained in trials with treatment-resistant major depressive disorder.