Abstract
The clinical and research literature on “treatment-resistant depression” is sampled and reviewed. Major conceptual and methodological inadequacies are identified in this literature. The main conceptual problem consists of a confounding of syndrome description with clinicians' treatment-outcome expectations. Methodological problems include: (a) the failure to adequately conceptualise and/or specifically define treatments; (b) the extreme variety of treatments evaluated in the absence of methodological controls; (c) inadequate sampling procedures; (d) the heterogeneity of the patient samples; (e) the bias in favour of reporting positive results; (f) the misrepresentation of outcome in published reports; and (g) the neglect of considering spontaneous remission as a viable alternative explanation. It is recommended that research attention be refocussed on defining the characteristics of the depressed person who does not respond to standard anti-depressant therapies, as a precursor to identifying a specific therapy.