Put the essay's two prevalence figures next to each other and the gap is easy to state. If about 4 percent of Americans with schizophrenia receive clozapine [3], and roughly a third of patients could benefit because their illness does not respond to other antipsychotics [4], then about 12 percent of the potentially eligible group gets the drug and close to seven in eight do not [1]. That ratio is what the registry was blamed for. It is also the quantity that nobody in this account has measured since the registry was retired in February 2025 [1].
The mechanism argument is clinical and first-person. The author, a child psychiatrist writing in STAT who ran controlled trials of clozapine in young people [8], describes watching capable colleagues reach for a fourth or fifth less effective medication because avoiding clozapine felt safer [9], and reports being unable recently to find a single child psychiatrist across four academic centers willing to start it in a child with childhood-onset schizophrenia [7]. That is a real observation about the supply of willing prescribers in one referral network. It falls short of a rate: the essay does not say how many clinicians were asked, over what period, or how those four centers compare with others.
There is a tension inside the piece worth naming. It holds that there is good evidence the monitoring program itself kept the drug from patients who needed it [6], and also that the paperwork was never really the problem [5]. Both can be true if the registry was a secondary barrier, which is the reading the author's own account supports. But that version makes a modest prediction, not a dramatic one, and the prediction is testable: initiations should rise a little after February 2025, not step-change.
The pediatric case is harder than the adult one on exactly the axis clinicians fear. Childhood-onset schizophrenia tends to be more treatment-resistant, which strengthens the argument for using the most effective drug [11], while neutropenia, the sharp drop in infection-fighting white cells the registry existed to catch [2], is more common in children than in adults [10]. So the group with the strongest benefit case also carries the higher rate of the adverse event driving avoidance. Any honest training push has to start there rather than around it.
The thing the essay does not tell you is whether familiarity actually changes prescribing behavior at scale. The author's fear resolved with experience [12], and parents in their practice arrived already persuaded and often ahead of the clinicians [13], but personal testimony falls short of a trial of an intervention. A mentoring or consultation service, measured against initiations per eligible patient rather than against surveyed clinician confidence, would test the claim. Until someone publishes that denominator, the removal of the registry is documented, while its effect on children with psychosis remains unmeasured.
