Evidence question
Screened adults with MDD, one 25 mg oral synthetic-psilocybin administration, monitored clinical setting, controlled comparison, and prespecified 24-hour and seven-day harms.
Independent replication package · framework v0.7
The protocol is frozen; the external-review launch gate is now explicit.
The question, search plan, forms, analysis rules, reconciled 13-record horizon corpus, and audit trail are packaged for a real independent team. This page still does not claim that independence has happened.
Complete1 estimand
Narrow clinical question and interpretation boundary frozen.
Prepared10 searches
Database, registry, citation, and grey-literature strategies specified.
Unassigned2 + 1
Two independent reviewers and one adjudicator still required.
Not complete0 registrations
PROSPERO/OSF submission requires an accountable human review team.
01 · Research state
A prospective replication of a disclosed retrospective pilot.
DoneAuthor pilot
Seed search, extraction, provisional synthesis, and certainty analysis.
DoneProtocol package
Estimand, eligibility, outcomes, forms, and analysis rules frozen.
NextExternal launch
Recruit the team, register prospectively, obtain PRESS review, then execute and archive searches.
PendingDual review
Concealed first-pass screening, extraction, and RoB 2.
FutureAdjudicated release
PRISMA flow, independent GRADE, frozen data, and rerun code.
Why this wording matters
OSF defines preregistration as a time-stamped, read-only plan posted before the work it governs. The existing author pilot cannot be made preregistered after the fact. The new registration can prospectively govern the independent replication while openly preserving the pilot as prior work.
02 · Protocol lock
The rules are fixed before new reviewers see the answer.
Strata
Direct MDD evidence remains separate from TRD, lower/weight-adjusted dose, repeated-dose, fixed-order, and incompatible-comparator evidence.
Rare-event math
Arm counts first; exact intervals; no continuity correction that creates information; no unstable relative-effect pooling when events are absent.
Interpretation
No recreational-use generalization, no “risk is zero,” no safety equivalence, and no conversion into a universal 0–100 drug score.
03 · Search architecture
The search is broad enough to find disagreement.
The new strategy searches from database inception rather than 2018 and adds psilocin, COMP360, broad depression terms, registries, citation networks, sponsors, and author contact. Eligibility—not a narrow query—does the final filtering.
| Source | Status | Purpose | Required proof |
|---|---|---|---|
| PubMed/MEDLINE | Pilot only | Core biomedical publications | PRESS-reviewed final query, date, count, NBIB/CSV export |
| Embase + PsycINFO + CENTRAL | Not run | Additional trials, psychiatric indexing, conference records | Licensed-platform histories and exports |
| ClinicalTrials.gov + WHO ICTRP | Partial | Unreported, ongoing, terminated, and discrepant studies | All statuses, record histories, second-person verification |
| Citation and grey literature | Not run | Linked reports, regulatory material, missing arm-level harms | Source chain, contact log, archived responses |
Every exact draft query is in search-strategies-v0.6.csv, with an accompanying peer-review form.
04 · Independent workflow
Independence is implemented as access control, not a promise.
First pass
Screens, extracts, and judges RoB 2 without seeing Reviewer 2 or the author-pilot decisions.
Independent first pass
Uses the same forms with concealed decisions. Uncertainty moves records forward rather than forcing exclusion.
Conflict resolution
Resolves only after both passes lock, cites the rule or source used, and preserves the disagreement.
Search peer review
Checks conceptual translation, headings, text words, syntax, filters, and database translations before execution.
05 · Candidate universe
Thirteen seed records are registered; none bypass screening.
The public horizon scan now includes the 2026 Yngwe MDD trial and EPIsoDE TRD trial, plus phase-3 and ongoing registry records. Seed records improve recall, but their presence does not predetermine inclusion.
| Candidate | Likely stratum | Why it matters | Independent status |
|---|---|---|---|
| Raison 2023 | Direct MDD | Single 25 mg, niacin control, supported setting | Awaiting dual screen |
| Yngwe 2026 | Direct MDD | Reports no drug-related SAE but severe anxiety requiring medical attention in two participants | Awaiting dual screen |
| Goodwin 2022 + Mertens 2026 | Indirect TRD | Important safety evidence from a different baseline-risk population | Awaiting dual screen/extraction |
| von Rotz, Davis, Sloshower, escitalopram trial | Indirect dose/regimen/design | Tests how conclusions change outside the exact exposure | Awaiting dual screen |
| uAspire and phase-3 registries | Future/direct or indirect | Prospective evidence and missing-publication surveillance | Awaiting results/full reports |
06 · Registration gate
The files are ready; the accountable research team is not.
PROSPERO
PROSPERO is an international prospective registry for health-related systematic reviews. A team member must create the record, take responsibility for the review, and keep it updated.
OSF
An OSF registration can archive a time-stamped, read-only version of the protocol, forms, code, and manifest. It should explicitly say that the registration governs the independent replication, not the completed author pilot.
Release condition: do not change the public label from “single-author pilot” until named independent reviewers complete the workflow and the adjudicated data/code release is posted.
07 · Launch gate
The handoff now has a hard operational boundary.
v0.7 corpus repair and launch controls
The candidate CSV, RIS, detailed corpus, family registry, and version metadata now agree on 13 seed records. Open the external-review launch page → for recruitment, search review, data management, and guarded packet-generation files.
07 · Replication files
A reviewer can inspect the plan without touching the website code.
Frozen protocolMarkdown · eligibility, outcomes, synthesisRegistration draftMarkdown · PROSPERO/OSF copy-ready fieldsSearch strategiesCSV · 10 sources and exact draft queriesSearch peer reviewCSV · PRESS-style checksScreening assignmentsCSV · concealed dual-review queueExtraction formCSV · arm-level harms and provenanceRoB 2 formCSV · result-level judgmentsGRADE formCSV · independent certainty decisionsTeam rolesCSV · qualifications and independence rulesProtocol lockJSON · SHA-256 file manifestPRISMA flowCSV · pending independent countsReviewer guideMarkdown · stage-by-stage instructions