Abstract
<jats:p>Antidepressants are widely prescribed, but their routine use remains contested. This critical narrative review evaluates efficacy, safety, and symptom-based prescribing using regulator-held and RIAT trial reanalyses, patient-level and variance-ratio meta-analyses, pharmacovigilance, and trial-methodology research. Pooled across all registered trials, the average drug–placebo difference is below two points on the 17-item Hamilton scale, beneath thresholds of clinical relevance for most patients, while selective publication, placebo washout, and broken blinding inflate apparent efficacy. The claim that this small average conceals genuine responders lacks support: parallel-group trials cannot estimate individual treatment effects, and four variance-ratio meta-analyses of more than 100,000 patients find no excess drug-arm variability. A modest effect on core mood is probably genuine, but evidence for a substantially benefiting subgroup is weak; in the largest pragmatic trial, about 3% sustained remission for one year. Harms—including suicidality in younger patients, akathisia, mania, persistent post-SSRI sexual dysfunction, emotional blunting, and a withdrawal syndrome routinely misclassified as relapse—have been understated. The critique is specific to depression; in children and adolescents efficacy is weaker still, whereas for anxiety disorders, obsessive-compulsive disorder, and premenstrual dysphoric disorder the evidence is stronger. Major depression is a symptom-defined syndrome and final common pathway for many treatable medical and psychosocial conditions; checklist-based prescribing may therefore treat a label, not its cause. Excluding treatable causes should precede prescription, and psychotherapy and cautious, time-limited prescribing deserve greater priority.</jats:p>