What never reached the doctors

Half the trials were positive. Almost all the published ones were positive.

Researchers compared 73 trials of 12 antidepressants (12,564 patients) registered with the FDA against what was eventually published. Only 50 (69%) were ever published — and the negative ones were the ones that disappeared.

The result: the therapeutic benefit, as it appeared in the published literature, was inflated by 32% overall — ranging from 11% to 69% depending on the drug.

[9] N Engl J Med 2008;358:252–260

The trial that reported the opposite of its own data

In 2001 a study funded by SmithKline Beecham (later GSK) presented paroxetine as effective and safe for adolescents with depression. It became one of the most influential papers in the field.

In 2015 an independent team repeated the analysis from the primary data, under the BMJ's RIAT initiative. The conclusion was the reverse: paroxetine did not differ from placebo on efficacy, while the group receiving it showed more suicidal ideation and behaviour.

Fourteen years passed during which the original version circulated as science.

In 2025–2026 the story gained one more chapter, fifteen years after the data were corrected. On 30 September 2025 JAACAP issued an Expression of Concern for the 2001 study, following a lawsuit from attorney George W. Murgatroyd III — the same attorney who obtained access to GSK's internal archives for the RIAT reanalysis above — against AACAP and Elsevier. On 24 March 2026 a judge in DC Superior Court dismissed the suit: ruling that a journal article is not a "consumer product" under the relevant consumer-protection statute — a procedural ground, with no assessment of the study's scientific accuracy.

Neither development changes the 2015 finding above, and neither one clears or convicts the 2001 study on its own: the dismissal is procedural, not a scientific verdict; the Expression of Concern is not a retraction and not a finding of fraud, it is a holding notice. As of our last check (September 2026) it remained published; no further action has been announced.

The authorship of the paper — who wrote the first draft, and for what fee — is examined on the sister site: Part E of psychpractices.org. Here we stay with the data.

[10] Le Noury J. et al. "Restoring Study 329", BMJ 2015;351:h4320 · [55] JAACAP, Expression of Concern, 30.09.2025 · Murgatroyd v. AACAP & Elsevier, Superior Court of the District of Columbia, No. 2025-CAB-005368, dismissed 24.03.2026

Retraction of an intravenous ketamine trial

On 19 January 2026, BMC Psychiatry retracted a 2024 study of intravenous ketamine for treatment-resistant depression with suicidal ideation, after editors themselves found discrepancies between the original data and what was published. Verbatim from the retraction notice: "The Authors provided some source data on request from the Editors, but an analysis by the Journal found significant discrepancies between this data and the data reported in the article... The authors did not explicitly state whether or not they agree with this retraction."

What this does not show. This concerns the data integrity of one trial from one research group — it is not an evaluation of ketamine's efficacy or safety more broadly, and it proves nothing about the doctor or clinic treating you today. Esketamine's (Spravato) approval rests on a separate body of studies unrelated to the retracted paper, and multiple other randomised trials of ketamine exist outside this one.

If you're reading this in crisis: the Suicide Intervention Line 1018 (Greece, KLIMAKA) runs 24 hours a day, free and anonymous.

[56] BMC Psychiatry, Retraction Note, 26:48 (19.01.2026) · doi:10.1186/s12888-026-07808-5 — original article BMC Psychiatry 2024;24:341 · doi:10.1186/s12888-024-05716-0