Phase 2 and Phase 3 ask different questions

Phase 2 explores whether a treatment shows useful benefit and which dose or regimen to pursue. Phase 3 aims to confirm benefit and safety in a broader, usually larger program suitable for a regulatory application. Sample size matters, but purpose, controls, follow-up and the strength of the evidence matter too. COMP005 and COMP006 are trial names; they do not mean Phase 5 and Phase 6.

MADRS: lower is better

MADRS is a clinician-rated depression symptom scale from 0 to 60. A negative change usually means symptoms fell. A treatment-control difference of −3.8 means the treatment group improved by an additional 3.8 points on the study’s adjusted analysis. Compare the same endpoint and timepoint. A larger number of participants, or the label COMP006, says nothing by itself about symptom worsening.

What a small p-value tells you

A p-value asks: if there were no group difference, and the statistical assumptions held, how unusual would results at least this extreme be? p<.001 means less than 0.1% under that model. It is not a 99.9% chance that the drug works, a measure of improvement size, or proof that bias is absent.

A percentage needs a denominator

GH’s 57.5% is 23 out of 40, not a precise promise for future patients. My Wilson 95% confidence interval calculation is approximately 42.2%–71.5% for that group; placebo 0/41 gives approximately 0%–8.6%. These are separate group intervals, not an interval for the difference. Repeated use of this interval procedure has about 95% coverage under its assumptions. It accounts for sampling uncertainty, not unblinding or selection bias. [3]

The five-line note I keep

Stage: what has actually finished? Evidence: how much, how fast, how durable, compared with what? Delivery: total visit time and resources? Cash: date, balance and guidance? Financing: gross or net proceeds, and how many new shares? Those five lines make each quarterly update comparable.

Sources

Plain-text references. Financial balances are dated June 30, 2026; source review completed October 6, 2026.

  1. Compass Pathways. Second Phase 3 COMP360 primary-endpoint announcement. February 17, 2026.
  2. Cubała WJ, Bajbouj M, Bauer M, et al. GH001 vs Placebo in Patients With Treatment-Resistant Depression: A Randomized Clinical Trial. JAMA Psychiatry. 2026;83(6):561–569. DOI: 10.1001/jamapsychiatry.2026.0096.