Reviews/Ssris and Snris/Antidepressant effect size is small versus placebo in mild-to-moderate depression
· SS-MDD-02 · v2
Borderline: within 4 points of the Likely/Uncertain boundary, so a re-grade could land either side.
Most of the usable studies here measured how much antidepressants beat placebo across all adults with major depression, and they agree the gap is modest: one patient-level analysis reports an effect size of around 0.30 [7], and another finds 0.33 to 0.48 even in a design built to favour the drug [16]. The largest dataset, covering 73,388 participants, puts the average drug-placebo difference at 1.75 points on a standard depression scale and finds that gap grows as patients start out more severely ill [1], which implies the effect is at or below that modest level at the milder end. The studies that focused specifically on mild or sub-threshold depression were reviews or opinion pieces rather than new trials, so they were set aside, leaving the mild-to-moderate reading resting mainly on that single severity pattern [1]. Several supporting papers come from the same research group and share overlapping trial data, so they are not fully independent. There is also a live argument that the standard scoring may understate the drug's true effect [13][16]. The panel treated the overall modest magnitude as well supported while treating the mild-to-moderate specificity as the weaker link.
Across the primary studies that survived filtering, the magnitude is consistently modest: [7] reports an effect size of around 0.30, and [16] reports 0.48 for citalopram and 0.33 for paroxetine even in a design built to maximise the drug signal. [1], the largest individual-participant analysis of 73,388 participants, records a mean drug-placebo difference of 1.75 HAMD points and shows the difference increases with baseline severity, so the effect at the milder end sits at or below that value. [13] found that 18 of 32 comparisons (56%) failed to separate drug from placebo on the HAMD-17 sum score. The direct link to the mild-to-moderate band comes chiefly from [1]'s severity gradient, because the mild-specific papers [8], [9] and [12] were reviews or editorials and excluded. The contested point: The division is whether a modest effect overall can be read as modest specifically in mild-to-moderate depression. That inference rests on [1]'s finding that the effect rises with severity, an association disputed elsewhere [20], while [13] and [16] argue the standard sum-score may understate the true drug effect.
All 20 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Ssris and Snris, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
Three of the four surviving primary papers ([7], [13], [16]) report overall adult-MDD effect sizes, not a mild-to-moderate subgroup; only [1] supplies the severity link. The genuinely mild-focused evidence ([8], [9], [12]) was filtered out as reviews, so direct primary evidence for the exact band is thin.
Evidential dependence: [7], [13] and [16] share the Hieronymus research group and overlapping FDA SSRI datasets, so they are not independent replications; effective independent weight is closer to [1] plus one cluster.
Where it remains uncertain. The mild-to-moderate reading depends mainly on [1]'s severity gradient, and [13] and [16] contend the HAMD-17 sum score understates the real effect.
Aleatory
[7] identifies a large and variable placebo response as a factor reducing the separation achievable, adding irreducible noise to each trial.
Epistemic
The mild and sub-threshold papers [8], [9] and [12] were reviews or editorials and excluded, so no primary study here measures the exact band.
Inferential
The reading relies on inferring the mild-to-moderate effect from [1]'s severity gradient rather than direct measurement of that band.
The load-bearing papers from a corpus of 20, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Largest FDA individual-participant dataset: small absolute separation overall and a smaller effect at lower (mild-to-moderate) baseline severity.
“The random effects mean difference between drug and placebo favored drug (1.75 points, 95% confidence interval 1.63 to 1.86). Differences between drug and placebo increased significantly (P<0.001) with greater baseline severity.”
Panel auditInternal validity: high · scope of inference: Adults and children in FDA-submitted acute antidepressant monotherapy efficacy trials, 1979-2016.Key limitations: mild-to-moderate stratum not isolated · trial populations may underrepresent very mild cases
States the headline standardised effect size, which is 'small' by conventional thresholds.
“Antidepressants outperform placebo with an effect size of around 0.30.”
Panel auditInternal validity: medium · scope of inference: SSRI placebo-controlled trial participants, theoretical ceiling of achievable effect sizes.Key limitations: 0.30 stated as background not stratified result · simulation-based · not mild-to-moderate specific
Patient-level SSRI effect sizes in the small range even in an analysis designed to isolate a genuine drug effect.
“the effect sizes being 0.48 for citalopram and 0.33 for paroxetine”
Panel auditInternal validity: medium · scope of inference: Adult major depression in citalopram/paroxetine trials, testing the side-effect-expectation hypothesis.Key limitations: single-item outcome · not severity-stratified · mechanism not magnitude focus
On the standard sum-score outcome a majority of comparisons do not separate, consistent with a small magnitude of effect.
“While 18 out of 32 comparisons (56%) failed to separate active drug from placebo at week 6 with respect to reduction in HDRS-17-sum”
Panel auditInternal validity: medium · scope of inference: Adults with major depression in SSRI industry trials, signal detection across rating measures.Key limitations: no effect size magnitude · not severity-stratified · single-item outcome is contested
4 of 20 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Small overall effect size is robustly supported, but the mild-to-moderate reading depends chiefly on [1] and the 'small' label is convention-dependent, holding confidence below Established.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Direct primary trial in mild-to-moderate MDD. The band the claim names is measured by no primary study here; the reading is inferred from an overall-population gradient. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A placebo-controlled trial enrolling only mild-to-moderate patients (e.g. baseline HRSD <=20) reporting a patient-level effect size.
Open question: What is the drug-placebo effect size when only mild-to-moderate patients are enrolled? Why it is still open: No primary trial in the corpus restricts to that band; the mild-specific papers [8], [9] and [12] are review-tier and excluded.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 26 Oct 2026 · protocol v0.19