The most common mistake in treating depression is thinking of resistance as a switch that flips after a second failed prescription. The specialists who gathered in Boston on December 6, 2025, for a consensus panel on mechanism-informed sequencing argued for a different image: resistance as a slope.
You do not suddenly become treatment-resistant. You slide.
The clinical convention is a binary: less than 50% improvement after two adequate antidepressant trials, and the label applies. That definition was built for regulators, not for patients. It is pragmatically simple, but it does not account for subtherapeutic dosing, insufficient duration, poor adherence, diagnostic uncertainty, or the influence of comorbid psychiatric and medical conditions.
The label sounds like a verdict. The reality is often a paperwork error, a missed diagnosis, or a mechanism that was never tried.
A patient can be labeled treatment-resistant without anyone knowing whether the resistance is biological or simply a series of mismatched, poorly delivered, or unexamined treatments. That is the contradiction at the center of the field. The label sounds like a verdict. The reality is often a paperwork error, a missed diagnosis, or a mechanism that was never tried.
The panel kept the term TRD for consistency with regulatory and clinical usage but reframed it as a continuum of increasing treatment complexity that evolves over time. Biological vulnerability, psychological processes, environmental stressors, and treatment-related factors interact dynamically, often through epigenetic mechanisms, shaping both the illness trajectory and the response to treatment. Resistance develops gradually as ineffective or poorly matched interventions accumulate.
The cost of that accumulation is measured in missed time. Early symptom improvement is a robust predictor of long-term outcome, yet in routine practice, ineffective treatments are frequently continued without structured reassessment. The panel defined early improvement as approximately 20% symptom reduction within two weeks of starting therapy and 25% to 50% reduction by four to six weeks on validated measures such as the Patient Health Questionnaire-9.
Those are not arbitrary numbers. They are a checkpoint. If a patient has not moved 20% in two weeks, the mechanism is probably wrong, the dose is probably wrong, or the diagnosis is probably incomplete. Waiting another six weeks for a formal failure is not caution. It is delay.
The first dimension worth measuring is anhedonia, the impaired ability to experience pleasure. It is not just a symptom.