is an 11-point madrs gain about four times what antidepressants usually beat placebo by?
“The difference here was huge, and 11 point change on the Madras MADRS is about four times what we see with antidepressants.”
a host, a psychiatrist said on The Carlat Psychiatry Podcast, 2026-11-02, at 15:23. hear it in context →
the episode: Kraepelin on Clancy Part 1 (at 15:23). we link the show's own page; we do not re-host audio.
first, the part that holds
the arithmetic holds. across 28 placebo-controlled trials of newer antidepressants that used the MADRS, drugs beat placebo by about 3 points (2.99), so an 11-point gap over placebo is about 3.7 times that. "about four times" is a fair rounding.
what we found when we went and looked
we found the typical antidepressant-versus-placebo gap on the MADRS in a meta-analysis of 28 randomized trials with 11,705 participants: 2.99 points (95% ci 2.24 to 3.74). 11 divided by 2.99 is 3.7, and against the ends of that interval it runs from 2.9 to 4.9 times. an FDA analysis of 232 trials on a different scale (HAMD) found a similarly small gap of 1.75 points, which backs up the "small typical effect" side of the comparison. the comparison only works if the 11 points is a drug-minus-placebo difference. in one antidepressant's FDA label, the drug groups improved by 13.0 to 20.4 MADRS points from baseline, so an 11-point within-group change would not beat antidepressants at all. the speaker named no study, so we could not check the 11 itself.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| the typical MADRS improvement with antidepressants versus placebo is roughly one quarter of 11 points (about 2.75 points) | confirmed | a meta-analysis of 28 placebo-controlled trials of new-generation antidepressants (11,705 participants) found a drug-placebo gap of 2.99 MADRS points (95% ci 2.24 to 3.74), with a standardized mean difference of 0.30. 11/2.99 = 3.7. limitation: this pools trials from one curated dataset, and single trials vary widely. in one antidepressant's label, individual-trial gaps ran from 1.5 to 7.1 points. [1][3] |
| an 11-point MADRS improvement is about four times the typical antidepressant effect | confirmed | this is true only when both numbers are drug-minus-placebo differences. 11 is 3.7 times the pooled 2.99-point gap. if the 11 were a raw change from baseline, the comparison fails: in the vortioxetine label, drug groups improved 13.0 to 20.4 points and placebo groups 10.8 to 14.5 points from baseline. [1][3] |
| the treatment discussed on the show produced an 11-point MADRS change | unverifiable | the speaker named no drug, trial or publication, so we could not find the record behind the 11-point figure. we also could not tell whether it was a placebo-subtracted difference or a change from baseline. the claim sentence treats the 11 points as given, so this row is context, not part of the verdict. |
| the antidepressant-placebo gap is small in the FDA's own trial archive | confirmed | an individual participant data analysis of 232 trials submitted to the FDA (73,388 participants) found drug improvement of 9.8 points, placebo improvement of 8.0 points, and a gap of 1.75 points (95% ci 1.63 to 1.86). this is on the HAMD, not the MADRS, so it supports the size of the typical effect but cannot be plugged directly into the 4x ratio. [2] |
both directions
for the speaker: the best pooled MADRS figure we found, a 2.99-point drug-placebo gap across 28 trials, makes 11 points about 3.7 times the typical effect. the FDA's own archive of 232 trials shows a small average gap of 1.75 HAMD points. the same meta-analysis says the typical MADRS gap falls below the roughly 8 points a clinician needs to see a minimal improvement. an 11-point gap over placebo would clear that bar, which fits the speaker calling it huge.
against: "typical" hides wide spread. in a single antidepressant's FDA label, individual-trial MADRS gaps ran from 1.5 to 7.1 points. against the best of those trials, 11 points is about 1.5 times, not four. the quote says "11 point change," which listeners could easily hear as change from baseline. on that reading the comparison is wrong: antidepressant groups in that label improved 13.0 to 20.4 points from baseline, and placebo groups 10.8 to 14.5. new treatments are often tested in small, early trials that may not be properly blinded, where effects tend to be larger than in the large trials behind the 2.99-point average. we could not check whether that applies to the trial behind the 11.
what this cannot say
we could not identify the trial or drug behind the 11-point figure, so we checked the comparison, not the result itself. the 2.99-point pooled figure covers newer-generation antidepressants in one curated trial dataset. older drugs, other populations and trials that did not use the MADRS are outside it. the FDA archive uses a different scale, and we did not convert HAMD points to MADRS points. averages of drug-placebo gaps say nothing about how much any one person improves. the FDA analysis itself found that about a quarter of people on drug had large improvements.
the medication pages behind this check
sources, each one fetched and read
- Hengartner MP, Jakobsen JC, Sørensen A, Plöderl M. Efficacy of new-generation antidepressants assessed with the Montgomery-Asberg Depression Rating Scale: a meta-analysis of randomised placebo-controlled trials. PLOS ONE, 2020. What we read there: MADRS: 28 trials, 11,705 participants; drug-placebo raw score difference 2.99 points (95% CI 2.24 to 3.74); SMD 0.30 (0.22 to 0.38); clinicians need about 8 MADRS points to detect minimal improvement. Fetched 2026-11-02 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC7043778/
- Stone MB, Yaseen ZS, Miller BJ, et al. Response to acute monotherapy for major depressive disorder in randomized, placebo controlled trials submitted to the US Food and Drug Administration: individual participant data analysis. BMJ 2022;378:e067606. doi:10.1136/bmj-2021-067606 What we read there: 232 trials, 73,388 participants; mean drug improvement 9.8 points, placebo 8.0 points, difference 1.75 points (95% CI 1.63 to 1.86) on the HAMD; 24.5% on drug vs 9.6% on placebo had large (about 16-point) improvement. Fetched 2026-11-02 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/9344377/
- vortioxetine tablets, prescribing information, section 14 clinical studies. DailyMed, US National Library of Medicine. What we read there: adult MDD trials: drug-group MADRS change from baseline -13.0 to -20.4, placebo -10.8 to -14.5; placebo-subtracted differences -1.5 to -7.1 points. Fetched 2026-11-02 (HTTP 200). https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=1a5b68e2-14d0-419d-9ec6-1ca97145e838
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