Despite major expansions of evidence-based treatments of common mental disorders in recent decades, especially antidepressant medication, the point prevalence of depression has not decreased; instead it probably increased in young adults. We question whether antidepressants (AD)-monotherapy and low-fidelity-to-guideline psychological treatment (PT) might have no effect or even adverse effects in some patients and contexts that dilute the benefits of treatment at the population level, making it harder for population-based studies to detect treatment-driven prevalence reductions. Randomized Clinical Trial (RCT)s have not identified these effects because AD-monotherapy and low-fidelity PT are uncommon in RCTs where treatment protocols are specified and carefully monitored, unlike treatment in real-world settings. Second, RCTs may have missed the bigger picture of ultimate outcomes due to too short follow-ups. We elaborate two mechanisms through which AD-monotherapy and low-fidelity PT could produce adverse effects on long-term illness course. Both mechanisms are speculative and we outline how to test. Major depressive disorder is a common, highly heterogeneous disorder. To reduce its enormous burden [1], Western countries have expanded their mental health expenditures in specialty and primary care settings since the 1980s [2], providing more treatment for more people, primarily with antidepressants (AD) and, to a lesser extent, psychological treatments (PTs) [3,4]. Both Antidepressant (AD)s and PTs are well-established, evidence-based treatments with roughly similar modest short-term efficacy. Surprisingly, the point-prevalence of depression in the general population has not decreased since the wide-scale use of AD [5]; instead, the prevalence has probably increased, particularly in young people [6]. In addition, the prevalence of recurrentchronic and treatment-resistant depression seems to have risen [7]. One obvious explanation for this "treatment-prevalence paradox" is increased first incidence, offsetting a treatment-driven prevalence drop. But population-based incidence studies do not find such an increase [4,8]. What else could explain the paradox? Explaining the Treatment-Prevalence Paradox We question whether AD-monotherapy and low-fidelity-to-guideline PT might have no effect or even adverse effects in some patients and contexts that dilute the benefits of treatment at the population level, making it harder for population-based studies to detect treatment-driven prevalence reductions. The hypothesized mechanisms through which AD-monotherapy and low-fidelity PT could produce adverse effects on long-term illness course include Loss of Agency [9] and Oppositional Perturbation [10] or Tolerance [11,12]. Both mechanisms are speculative and need rigorous testing. RCTs have not identified these effects because AD-monotherapy and low-fidelity PT are uncommon in RCTs where treatment protocols are specified and carefully monitored, unlike treatment in real-world settings. In addition, AD-withdrawal studies may have m...