Psilocybin Versus Escitalopram: Why the Primary Endpoint and Remission Rates Tell Different Stories

Can one depression trial produce a nonsignificant primary comparison and still report more remissions in one group? Yes. Those results can coexist because they summarize different features of the data and carry different statistical safeguards. The original psilocybin-versus-escitalopram trial offers a useful example of why comparative evidence cannot be reduced to whichever number appears most persuasive.

This article concerns the randomized six-week comparison, not the site's coverage of the later six-month follow-up. Its distinct question is how to interpret the original outcome hierarchy. The central issue is not whether a finding is interesting, but which conclusions the design and analysis can support about comparative depression outcomes.

Begin with the question chosen before the results

Carhart-Harris and colleagues randomized 59 adults with long-standing, moderate-to-severe major depressive disorder to two treatment regimens. Both groups received psychological support. The primary endpoint was change from baseline to week six on the self-reported Quick Inventory of Depressive Symptomatology, abbreviated QIDS-SR-16. That prespecified endpoint is the anchor for the trial's main efficacy comparison. [1]

Mean changes were minus 8.0 points in the psilocybin group and minus 6.0 in the escitalopram group. The between-group difference was minus 2.0 points, with a 95 percent confidence interval from minus 5.0 to 0.9 and a P value of 0.17. The primary comparison therefore did not establish a statistically significant difference between the regimens on that endpoint at six weeks. [1]

What that result does not establish

A nonsignificant result does not prove that the treatments are identical. The confidence interval describes uncertainty around the estimated group difference and includes both a potentially meaningful advantage for psilocybin and little difference. The observed two-point gap should not be treated as the exact population effect, nor should crossing zero be interpreted as evidence that no difference could exist.

Equivalence and noninferiority require designs and decision rules built for those claims, including an explicit margin. A trial that fails to demonstrate superiority has not automatically demonstrated either. Saying the treatments “worked equally well” therefore overstates this result. A more faithful statement is that this small trial did not establish a difference on its primary endpoint and left substantial comparative uncertainty.

Why remission can look more dramatic

The report also presented secondary response and remission outcomes. Response, defined as at least a 50 percent reduction in the depression score, occurred in 70 percent of the psilocybin group and 48 percent of the escitalopram group. Remission, using the study's symptom threshold, occurred in 57 percent and 28 percent, respectively. These are observed trial proportions, not personal probabilities for someone considering treatment. [1]

A continuous change score and a threshold outcome compress information differently. Two participants can improve by similar amounts yet land on opposite sides of a remission cutoff because they started at different severities or ended close to the boundary. Thresholds can make a clinically understandable outcome, but they also discard some information about the overall distribution. A larger-looking percentage gap need not contradict a less decisive primary analysis.

Outcome hierarchy prevents selective storytelling

The trial included multiple secondary outcomes, and those comparisons were not corrected for multiple testing. Testing many endpoints increases the opportunity to find an apparently favorable result by chance. Secondary findings can identify patterns worth investigating, especially when they are consistent, but they should not silently replace the primary outcome as the trial's definitive success criterion. [2]

This is why “psilocybin doubled remission” is incomplete as a standalone headline. It foregrounds a secondary proportion without communicating the primary finding or the multiplicity limitation. The opposite shortcut, “the trial showed nothing,” is also inaccurate. The study produced useful estimates and promising secondary signals. The responsible interpretation holds both facts together rather than discarding whichever one complicates the preferred story.

The comparator is a regimen, not every possible use

The study compared particular schedules of psilocybin and escitalopram within a supported research setting. It did not compare psilocybin with every antidepressant, every psychotherapy, or individualized long-term prescribing. Six weeks is also a defined observation period rather than a complete account of treatment continuation, relapse prevention, or care after an initial inadequate response.

Both the small sample and the difficulty of masking subjective drug effects constrain interpretation. People who infer their assignment may experience different expectations or report symptoms differently. The site's discussion of treatment guesses explains why support and randomization do not eliminate every measurement problem. These limitations do not erase the result, but they limit how confidently it can be generalized.

Compare benefits and burdens without inventing a winner

The original report stated that adverse-event incidence was similar between groups, but a small trial cannot establish that their complete safety profiles are equivalent. Different interventions can also entail different practical burdens, including supervised time, ongoing medication management, or uncertainty about durability. None of those considerations can be settled by selecting one depression percentage from the paper.

For a balanced summary, name the primary endpoint and its uncertainty first, then present response and remission as secondary findings with their analysis caveats. Finally, identify the trial's population, duration, support, and masking constraints. The result is a more precise conclusion: the original comparison generated encouraging signals for psilocybin, but did not establish superiority on its prespecified primary depression outcome or equivalence between the treatments.

Disclosure: It provides research education, not medical advice or a recommendation to use psilocybin.

Editorial disclosure: Generated pictures are illustrative, not study participants. Consult a qualified professional for personal care.