did 67% to 82% of catie patients quit their antipsychotic because it failed or because of side effects?
“Between 67% and 82% of the patients, depending on which drug they were on, dropped out of the trial because either the drug wasn't working or they couldn't stand the side effects.”
a guest, a historian of psychiatry (Andrew Scull) said on Lex Fridman Podcast, 2026-09-20, at 19:40. hear it in context → the speaker cited: the CATIE study, published in the New England Journal of Medicine.
first, the part that holds
The trial is real and the scale is close: in the NIMH-funded CATIE trial, published in the New England Journal of Medicine in 2005, 74% of patients stopped their first assigned antipsychotic within 18 months. By drug, the rates ran from 64% to 82%, so the top of the speaker's range is exactly right.
what we found when we went and looked
The 64–82% figures count every reason a patient stopped the first drug, not just the two the speaker named. In the papers we read, 24% of the 1,432 phase 1 patients stopped because the drug wasn't working and 10% to 18% (depending on the drug) stopped because of side effects. The largest single group, 30% (428 of 1,432), stopped by their own decision, with no further reason recorded. So "not working or side effects" covers about half of those who stopped, not all of them. "Dropped out of the trial" is also wrong for most of them: about 57% of those who stopped switched to another drug and stayed in the study. About 43% left it.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| between 67% and 82% of patients, depending on the drug, stopped their assigned antipsychotic | overstated | Secondary summaries of the primary paper give all-cause discontinuation rates of 64% (olanzapine) to 82% (quetiapine) over 18 months, and 74% overall. The top end matches. The bottom end is 3 points too high, so the best-performing drug looks a little worse than it was. We could not open the journal's own table, so the per-drug numbers come from peer-reviewed papers quoting it. [1][2][4] |
| those patients stopped because the drug wasn't working or they couldn't stand the side effects | misattributed | The 64–82% figure is all-cause discontinuation. Of 1,432 phase 1 patients, 24% (340) stopped because the drug wasn't working and 10% (risperidone) to 18% (olanzapine) because of side effects. The largest single group was patients' own decision to stop: 30% (428). Some investigator-initiated stops are also included. Not working plus side effects is roughly half of all stops, not the whole figure. Some 'patient decision' stops may have been driven by side effects or poor response, but the record does not show that. [1][3] |
| the patients dropped out of the trial | misattributed | Stopping the first drug was not the same as leaving the study. Under the design, patients could switch to another drug in later phases. Of those who stopped their first drug, about 43% left the study and about 57% switched treatment and continued in phase 1B or phase 2. [2][3][5] |
| the source is the NIMH-funded CATIE trial, published in the New England Journal of Medicine | confirmed | NIMH describes CATIE as its own trial and documents its phase 1 results. Papers citing it give the primary report as Effectiveness of antipsychotic drugs in patients with chronic schizophrenia, N Engl J Med 2005;353(12):1209–23. [1][3][4] |
| context: overall discontinuation and timing | confirmed | 74% of patients stopped their first assigned drug before 18 months, with a median of about 6 months. Median time to discontinuation was 9.2 months on olanzapine, 5.6 months on perphenazine, 4.8 months on risperidone, 4.6 months on quetiapine and 3.5 months on ziprasidone. [1][4] |
both directions
for the speaker: The core picture holds. Three in four people with chronic schizophrenia in a large, publicly funded trial stopped their first assigned antipsychotic within 18 months, most within about six months. The top of the speaker's range (82%) matches the record. Lack of effect (24%) and side effects (10–18% by drug) were both common reasons. Even the best-performing drug was stopped by nearly two-thirds of the patients on it. The trial was funded by NIMH and the primary report appeared in the New England Journal of Medicine, as the speaker said.
against: The figure the speaker quoted is all-cause discontinuation, the trial's composite primary outcome, and he presented it as if every stop was due to lack of effect or side effects. In the record, the single largest group (30% of patients) stopped by their own decision, with no further reason recorded, and side effects alone accounted for only 10–18% by drug. Most people who stopped their first drug did not drop out of the trial: about 57% switched to another drug and kept going. The low end of the range is 64%, not 67%.
what this cannot say
The journal's full text and the PubMed record were blocked to our fetcher. The numbers above come from NIMH's own results page and from peer-reviewed papers that quote the primary report directly, not from the primary report's tables. We did not read a per-drug breakdown of the 'drug wasn't working' stops, only the pooled 24%. The 'patient decision' category was not broken down further, so an unknown share of it may reflect side effects or poor response. The true share driven by those two reasons could therefore be higher than the recorded half, but the record cannot show that. The trial enrolled people with chronic schizophrenia over 18 months, so its stopping rates cannot be applied directly to first-episode patients, other diagnoses or other drugs. None of this is advice about anyone's own medication.
sources, each one fetched and read
- Effectiveness of antipsychotics: is the CATIE trial a tsunami? Canadian Family Physician, 2007;53(1). PMC1952563 What we read there: "74% of patients stopped treatment, with median time of 6 months"; "A quarter of the patients discontinued their antipsychotic treatment early due to inefficacy (24%, or 340 of 1432 patients), surpassed only by those discontinuing during phase 1 because of independent patient decisions to stop treatment (30%, or 428 of 1432)"; intolerability discontinuation ranged from 10% (risperidone) to 18% (olanzapine); "Patients also had the option of moving on to the second stage of the trial after discontinuation." Fetched 2026-10-13 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC1952563/
- Patient perception of medication benefit and early treatment discontinuation in a 1-year study of patients with schizophrenia. Patient Preference and Adherence, 2007;1:9–17. PMC2779124 What we read there: "In the CATIE study of 1493 patients with schizophrenia, patients randomized to 5 different antipsychotic drugs had all-cause discontinuation rates from 64% to 82% over an 18-month period"; "approximately 43% dropped out of the study opposed to approximately 57% who switched treatments and entered Phase 1b or Phase 2." Fetched 2026-10-13 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC2779124/
- National Institute of Mental Health. Questions and answers about the NIMH CATIE study: phase 1 results What we read there: "Almost three quarters of patients switched from their first medication to a different medication." Medications were stopped if they "did not control symptoms, or if the side effects were not tolerable, or if the patient chose to stop treatment for some other reason." Fetched 2026-10-13 (HTTP 200). https://www.nimh.nih.gov/funding/clinical-research/practical/catie/phase1results
- Time to discontinuation of first- and second-generation antipsychotic medications in the treatment of schizophrenia. Schizophrenia Research, 2011;131(1-3):127–132. PMC5935444 What we read there: "Seventy-four percent of individuals participating in CATIE discontinued their assigned medication before 18 months"; median time to discontinuation was 9.2 months on olanzapine vs 4.6 on quetiapine, 4.8 on risperidone, 5.6 on perphenazine and 3.5 on ziprasidone. Cites the primary report as N Engl J Med 2005;353(12):1209–1223. Fetched 2026-10-13 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC5935444/
- Interpreting and applying the CATIE results: with CATIE, context is key, when sorting out phases 1, 1A, 1B, 2E, and 2T. Psychiatry (Edgmont), 2007;4(10):23–29. PMC2860523 What we read there: Describes the multi-phase design: patients who stopped their phase 1 drug either left or entered phase 1B or phase 2 (2E or 2T) on a different drug. Phase 2E enrollees had stopped earlier phases "because of inadequate therapeutic effect (86%) or other reasons (14%)." Fetched 2026-10-13 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC2860523/
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam