In Reply Fogelson and Leuchter have responded to our proposed staging scheme for treatment-resistant depression (TRD) published in JAMA Psychiatry. 1 We do not fundamentally disagree with their arguments; however, we stand by our proposal that psychiatry needs to move toward a more empirically based TRD definition.Fogelson and Leuchter express concern about the lack of evidence to support using different antidepressant classes to improve antidepressant efficacy. In putting forward our preliminary TRD staging model, we included both psychopharmacology and psychotherapy treatments. Further, we describe preliminary treatment recommendations for stage I TRD, which include nonpharmacological treatments (electroconvulsive therapy and repetitive transcranial magnetic stimulation) as possible treatment trials for stage I TRD.If the treatments are pharmacological, the rationale for requiring failure of 2 different classes is based on existing limited evidence of the predictive value of specific antidepressant trial combinations in predicting future resistance. For example, does failure to benefit from escitalopram and citalopram (same class) convey the same information about future likelihood of benefit as a failed response to 2 medications from different classes, say escitalopram and nortriptyline? We believe there is no compelling evidence to address this question. When suggesting an arbitrary threshold and general guideline for defining TRD, it seems prudent to require 2 different treatment classes, whether within pharmacological interventions or across the different types of intervention.Existing data suggest the opposite problem. Evidence from a community referral TRD clinic found patients with TRD failed an average of 3.6 selective serotonin reuptake inhibitor class antidepressants, with many failing 5 selective serotonin reuptake inhibitors. 2 Thus, many patients with TRD fail a multitude of similar mechanism antidepressants for years.Regarding their other arguments (TRD is likely multifactorial/heterogeneous and TRD models based on number of failed treatment trials would not address subtypes): the authors suggest that patients with TRD and comorbid personality disorder or history of childhood adversity/trauma may require specific antidepressant interventions to be considered resistant. We agree that in the future, TRD staging will need refinement if clinical features/biomarkers that predict outcome with specific antidepressant interventions are determined. However, presently, we do not have this level of knowledge and these associations are not sufficiently robust or reliable to justify the refinements suggested by the authors.We concur with the ambitious goal of refining specific biomarkers and clinical TRD subtypes. However, until achieved, the severe costs of TRD treatment failure (10% per annum response rate to standard treatments 3 and high suicide risk 4 ) warrant a more practical and utilitarian TRD staging system to allow research and clinical practice to progress.Finally, Fogelson and Leuchter suggest th...