checkedThe Carlat Psychiatry Podcast2026-11-02last verified 2026-11-02

did madrs fall 11 points vs 4 when people with depression cut ultra-processed food?

verdict: misattributed
a real number attached to the wrong thing
“When they cut ultra-processed foods, depression fell 11 points on the clinician-rated mattress compared to four points while on their usual diet.”

a host, a psychiatrist said on The Carlat Psychiatry Podcast, 2026-11-02, at 14:41. hear it in context → the speaker cited: a pilot trial by UCSF researchers published in JAMA Psychiatry.

the episode: Kraepelin on Clancy Part 1 (at 14:41). we link the show's own page; we do not re-host audio.

first, the part that holds

the trial is real and the speaker has the main facts right. it is a pilot crossover trial led by ucsf researchers, published in jama psychiatry on september 23, 2026, and on the madrs depression scale people did much better while cutting ultra-processed foods than on their usual diet. the gap the paper reports is about 11 points (−10.78).

what we found when we went and looked

we read the trial's full text on the jama psychiatry site. 20 adults with major depressive disorder were enrolled and 19 finished an open-label crossover, with the order of the two diets randomized. the 11 in the quote is real, but it is the model-estimated difference between the two diet periods (−10.78 madrs points, 95% ci −13.90 to −7.66), not how far scores fell during the low-ultra-processed-food period. the published means were 21.6 at baseline, 8.05 in the low-ultra-processed-food period and 18.9 on the usual diet. that is a fall of roughly 13.6 versus roughly 2.7 points from baseline, and we found no 4-point figure in the paper. so the speaker's pairing understates the contrast the paper reports (a gap of about 7 points instead of about 11) while attaching the 11 to the wrong quantity.

assertion by assertion

assertionverdictwhat the record says
the study was a pilot crossover trial by ucsf researchers published in jama psychiatryconfirmedjama psychiatry, published online september 23, 2026: 'reducing ultraprocessed foods from the diets of patients with depression: a pilot crossover trial'. the lead author is in the ucsf department of psychiatry and behavioral sciences, with co-authors at the mayo clinic, the university of queensland and emory. the design was open-label and order-randomized, with 20 enrolled and 19 completing. [1]
madrs fell 11 points during the ultra-processed-food-reduction periodmisattributed−10.78 is the treatment difference between the two conditions, from a linear mixed-effects model adjusted for baseline mood score, period and sequence. it is not the fall within the low-ultra-processed-food period. the published means (21.6 at baseline, 8.05 in the low-ultra-processed-food period) put the fall from baseline at about 13.6 points.
madrs fell 4 points on the usual dietunverifiablethe full text we read gives no 4-point figure and no reported within-period change for the usual diet. its means (21.6 at baseline, 18.9 on the usual diet) imply a fall of about 2.7 points, which does not match 4. the 4 may come from a press account or from per-period starting scores we could not see.
implied gap between conditions (11 vs 4, i.e. about 7 points)undersoldthe paper's own between-condition difference is −10.78 madrs points (95% ci −13.90 to −7.66, p<.001), larger than the roughly 7-point gap the speaker's pairing implies. this is the exploratory result of an unblinded pilot with 19 completers.
the comparison period was 'ultra-processed-food-free'overstatedultra-processed food intake was reduced, not eliminated. the self-reported score fell from 10.10 to 1.42, about 86%, and the paper's title says 'reducing'.

both directions

for the speaker: the attribution checks out: ucsf-led, jama psychiatry, a pilot crossover with randomized order. depression scores were clearly lower in the low-ultra-processed-food period, on all three depression measures: madrs −10.78, the 9-item patient health questionnaire −3.12, and the 16-item quick inventory of depressive symptomatology −3.67, all p<.001. anxiety, on the 7-item generalized anxiety disorder scale, improved too (−2.74, p=.004). the direction and rough size of the madrs effect the speaker describes match the paper, and the paper's between-condition gap is larger than the speaker's framing implies.

against: the 11 is a difference between conditions presented as a within-period fall, and the 4 does not appear in the paper we read. the diet was a reduction, not a removal. the authors themselves frame the trial as a feasibility pilot whose mood outcomes are exploratory. it was nonblinded, intake was self-reported, and participants knew which diet they were on. that matters for any rated depression score, clinician-rated or not. about 65% of participants were taking psychiatric medication, though they agreed not to change it during the study. the authors conclude only that the pilot 'provides the basis for carrying out necessary future studies that are larger, blinded, use objective measures of upf measures, and assess mechanisms.'

what this cannot say

19 completers, open-label, a single site. the madrs results are exploratory, so they cannot establish that cutting ultra-processed food treats depression. on our side, the pubmed record has no abstract and the full text was read through a page-rendering tool, so we may have missed parts of the article or its supplement. we could not confirm the length of each period, whether the madrs raters were clinicians or blinded, or any per-period change scores. the 4-point figure may exist in material we did not see, which is why that row is unverifiable rather than false. we did not read the trial registration because its host is outside our allowed sources.

sources, each one fetched and read

  1. Reducing Ultraprocessed Foods From the Diets of Patients With Depression: A Pilot Crossover Trial. JAMA Psychiatry, 2026. PubMed record 42776545 What we read there: no abstract available. it is described as a pilot crossover trial assessing the feasibility and acceptability of removing ultraprocessed foods from the diets of patients with major depressive disorder. Fetched 2026-11-02 (HTTP 200). https://pubmed.ncbi.nlm.nih.gov/42776545/

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