The guidance and the ground-level reality do not match. In "Back to Basics: The Latest on Schizophrenia Spectrum Disorders," a 2026 SAMHSA-backed review, psychotherapy is treated as a core component of care, with cognitive behavioral therapy for psychosis the most-used approach and acceptance and commitment therapy, cognitive remediation, social skills training, and family models rounding out the list. Then there is what actually happens. In a prospective study across 15 Italian centers, of 116 people in the first five years of schizophrenia, all clinically stabilized on long-acting injectable antipsychotics, only 43 — 37.07% — received even one evidence-based psychosocial intervention.
Last issue, this space looked at how the field is redrawing the diagnostic boundary around schizophrenia. This week the question moves downstream, from classification to delivery: who gets treated, when, and with what.
The history explains why the delivery side got neglected. Nina Schooler, a social psychologist whose research career spans more than six decades, arrived at the National Institute of Mental Health's Psychopharmacology Service Center in the early 1960s, just as chlorpromazine reached the United States. Her boss, Jonathan Cole, scoured NIMH's grant portfolio for clinical psychopharmacology and found almost nothing, so he built a study instead: three phenothiazines against placebo, run across nine hospitals.
The site selection was the tell. Cole recruited state hospitals, acute admission facilities, and what were then called silk-stocking hospitals — Payne Whitney, the Institute of Living — where staffing was high, stays were long, and patients "actually received psychoanalytic therapy, which was at the time sort of the gold standard." The field's live question was whether antipsychotics worked at all, or only worked where treatment was otherwise poor.
The drugs won that argument, and the psychosocial side spent decades being downgraded to optional. The evidence never supported that downgrade. A 2013 review in the Annual Review of Clinical Psychology had already classified assertive community treatment, CBT for psychosis, cognitive remediation, family psychoeducation, illness self-management training, social skills training, and supported employment as evidence-based practices. Thirteen years on, the SAMHSA paper is still making the same case, which is itself the finding.
The Italian data sharpen what is at stake. Both the group that got psychosocial interventions and the group that did not improved on core symptom domains over 12 months. The difference showed up in real-world psychosocial functioning, where the intervention group was superior at 12 months and at every earlier time point measured.
That split matters clinically. Symptom control is what medication reliably delivers; functioning — work, relationships, daily life — is what patients actually live inside, and it did not follow automatically from symptom improvement.
Symptom control is what medication reliably delivers; functioning — work, relationships, daily life — is what patients actually live inside, and it did not follow automatically from symptom improvement.
Early intervention has the same shape: strong rationale, thinner execution. Coordinated specialty care, a team-based model combining psychotherapy, medication management, family services, case coordination, and supported employment and education, is now considered the standard of care for early psychosis by the American Psychiatric Association. It expanded from federal funding for 12 programs in two states in 2008 to 431 programs across every US state as of 2023.
The outcomes are real but bounded. In the RAISE-ETP trial, 33% of patients receiving coordinated specialty care were hospitalized for psychosis within two years. A meta-analysis of 10 trials comparing early intervention services with usual care, covering more than 2,000 participants, found 32% hospitalized within an average 16 months of follow-up.
Two risk factors drive that: residual positive symptoms and poor attitudes toward medication. Adherence data are brutal. Among 2,588 first-episode patients, only 1,507 — 58.2% — collected an antipsychotic prescription in the first 30 days after hospital discharge, and just 1,182, or 45.7%, stayed on that initial treatment for 30 days or longer.
Long-acting injectables help, though not uniformly. A cluster randomized trial of 489 first-episode and early-phase patients comparing injectables with clinician-choice medication produced a number needed to treat of 7 for preventing hospitalization. Other controlled trials have not shown the advantage, and methodologic problems limit how far the negative findings can be read.
The more useful point may be diagnostic. One-third of patients referred to a treatment-resistance program had subtherapeutic or undetectable blood antipsychotic levels, which is why the TRIPP working group recommends a long-acting injectable trial before anyone is labeled treatment-resistant. Roughly 15% to 20% of first-episode patients do meet criteria for treatment resistance early, and clozapine — the only antipsychotic approved for that subgroup — tends to arrive years late.
Now the boundary test. This October, The Lancet Psychiatry published a prospective, single-centre, assessor-blinded, parallel-group randomized controlled trial from Charité in Berlin on modified psychodynamic psychotherapy for schizophrenia-spectrum disorders. A generation ago, psychodynamic treatment was the establishment that the phenothiazine trials displaced; the school Schooler described as occupying the high-staffing hospitals. That a trial of this kind now runs with randomization and blinded assessment, and clears the bar for a journal of that tier, is the signal worth reading, whatever the effect sizes turn out to be.
Early intervention has its own costs, and the field is no longer pretending otherwise. Detection programs report false-positive rates for non-cases as high as 60%, and self-stigmatization has been documented in up to 25% of a sample of people with early psychosis. Disclosing risk can harm; withholding it can prolong untreated psychosis and override autonomy. There is no clean side.
Duration is the other unresolved problem. Most coordinated specialty care programs run about two years, a few stretch to three or five. Gains made during that window have often been lost when patients transition to usual care. A 10-year follow-up of an early intervention program outside the US found better employment outcomes than standard care — and still, 32% of the early intervention sample sat in a poor-employment cluster, with improvements sliding back to standard-care levels after five years.
So the shift underway is not a verdict on whether psychotherapy works for schizophrenia. That question was largely settled and then ignored. The shift is toward asking whether anyone receives it, how long care lasts, and whether functioning — not just symptom scores — is the outcome being measured.
The Italian study answers part of that. Psychosocial intervention predicted better real-world functioning in ordinary clinical settings, and its dissemination was still limited. The gap between the recommendation and the 37% is now the field's central problem, and it is a delivery problem, not a discovery problem.