checkedThe Carlat Psychiatry Podcast2026-11-09last verified 2026-11-09

did residents catch psychotic depression more often than attendings in the missed-diagnosis study?

verdict: unverifiable
no method, no file, or the data cannot answer it
“The residents in that study detected psychosis better than their attendings did.”

a host, a psychiatrist said on The Carlat Psychiatry Podcast, 2026-11-09, at 11:18. hear it in context → the speaker cited: one study (unnamed).

the episode: Kraepelin on Clancy Part 2 (at 11:18). we link the show's own page; we do not re-host audio.

first, the part that holds

the study is real, and it was designed to ask this question. a 2008 paper on missed psychotic depression at 4 academic medical centers listed the physician's level of training as one of the three factors it set out to test, and it found psychotic depression went undiagnosed in 27% of 130 diagnoses.

what we found when we went and looked

the speaker is talking about a 2008 paper in the Journal of Clinical Psychiatry. it covered 65 inpatients enrolled in the NIMH study of drug treatment for psychotic depression. by the study's research assessment, psychotic depression had been missed in 27% of 130 clinical diagnoses. the abstract on PubMed says the authors set out to "explore whether presenting symptoms, treatment setting, and physician's level of training affect the accuracy of diagnosis." its results section reports only symptoms and setting (inpatient units more accurate than emergency rooms, chi-square 7.64, p < .01). it gives no resident-versus-attending result. the full text holds that comparison, if it is reported anywhere. the full text sits behind the journal's paywall and we could not read it in this session, so we cannot confirm the claim or rule it out.

assertion by assertion

assertionverdictwhat the record says
there is a study of missed psychotic depression, and it found psychotic depression often goes undiagnosedconfirmeda 2008 paper in the Journal of Clinical Psychiatry (69(8):1293-6) found psychotic depression had not been diagnosed before research assessment in 27% of 130 diagnoses among 65 inpatients at 4 academic medical centers. the record we read gives the 27% figure; it does not settle the residents comparison. [1][2]
in that study, resident physicians detected psychosis more often than attending physicians didunverifiablethe abstract names physician's level of training as a variable the study explored. its results section reports no finding by training level: no rates, no test, no direction. the comparison, if reported, is in the full text, which is paywalled and which we could not read. a search-engine summary said attendings missed the diagnosis in 31 of 90 cases (34%) and that residents did better. we could not trace that to anything we read, so we do not rely on it. [1]
where the study did find a difference in accuracy, it was by treatment settingconfirmedthe abstract reports greater diagnostic accuracy on inpatient units than in emergency rooms (chi-square 7.64, p < .01). missed diagnoses were also more likely when depressed mood, hallucinations or delusions were not written in the chart (all p < .005). these are the only accuracy results the abstract reports. [1]

both directions

for the speaker: the study measured what the speaker describes. its stated objective includes testing whether the physician's level of training affects diagnostic accuracy. the count of 130 diagnoses for 65 patients fits the study collecting more than one physician's diagnosis per patient, which is what a resident-versus-attending comparison needs. that last point is our inference from the numbers, not something the abstract states. we also saw a search-engine summary that pointed the same way as the speaker (attendings missing about a third of cases, with residents' exposure to teaching on psychotic depression offered as an explanation). we could not read the text behind it, so it is not evidence here.

against: the abstract's results section, which is the public record of what the authors chose to report, never says residents did better. it reports symptoms recorded in the chart and treatment setting as the factors tied to missed diagnoses. if the training-level difference had been large and statistically significant, it would be odd to leave it out of the abstract. its absence is consistent with a weak, non-significant or descriptive-only difference, though it does not prove one. the abstract also does not say whether residents and attendings saw the same patients in the same settings. with emergency-room diagnoses already shown to be less accurate, any resident-attending gap could reflect where each group saw patients rather than training.

what this cannot say

we read the PubMed abstract and a later review that cites the study. we did not read the full article. it is paywalled on a publisher host outside our allowed sources, and our attempts to reach it in this session returned only a purchase page. so we cannot report the residents' and attendings' miss rates, the denominators, or whether any difference was statistically tested. the study itself is small: 65 patients at 4 academic centers in one NIMH treatment trial, from 2002 to 2004. even if residents did better in this sample, it may not generalize to community settings, to today's training, or to other levels of experience. the study's reference standard was a research diagnostic assessment, so a "missed" diagnosis means disagreement with that assessment.

sources, each one fetched and read

  1. Missed diagnosis of psychotic depression at 4 academic medical centers. J Clin Psychiatry. 2008 Aug;69(8):1293-6. doi:10.4088/jcp.v69n0813. PMID 18384244 (PubMed abstract record) What we read there: objective: "explore whether presenting symptoms, treatment setting, and physician's level of training affect the accuracy of diagnosis." results: "Psychotic depression had not been diagnosed prior to research assessments for 27% of the 130 diagnoses in our sample... The accuracy of diagnoses was greater on inpatient units than in emergency rooms (chi(2) = 7.64, p < .01)." the record contains no resident or attending result. Fetched 2026-11-09 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/18384244/
  2. Treatment for Major Depression With Psychotic Features (Psychotic Depression). Focus (American Psychiatric Publishing), 2016. PMC6519655 What we read there: cites the 2008 four-center study as finding "the diagnosis of psychotic depression was missed 27% of the time"; gives no breakdown by clinician training level. Fetched 2026-11-09 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC6519655

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