Peer Review Constitution / Pilot operating policy
A higher standard
of scrutiny.
The platform enforces a defined workflow. Editorial judgment remains accountable—and scientific certainty is never implied by publication.
Qualified before randomized
Editors verify reviewer credentials and approve declared topic expertise and review disciplines. The server excludes the submitting author, the assigning editor, previously invited reviewers in the same round, exact-match institutional conflicts, declared author/conflict email matches, unavailable reviewers, and reviewers at capacity.
Balanced contextual rotation
Within each required discipline, reviewers with fewer active invitations or reviews are prioritized, followed by those least recently assigned. Cryptographic randomness breaks ties. The algorithm version, eligible pools, selected reviewers, screening rationale, and time are recorded for editors. Authors cannot nominate reviewers.
Three complementary perspectives
Each round requires domain, methods, and clinical review. If the qualified pool is insufficient, the system stops rather than assigning an unqualified reviewer. Declined invitations can be replaced through the same selection process.
Conflicts require human scrutiny
The software screens declared structured information; it does not independently verify co-authorship, grants, commercial connections, mentorship, or institutional aliases. Editors must document that screening before assignment. Reviewers must affirm no conflicts before accessing the manuscript, or decline with a reason.
Double-anonymized by workflow
Authors upload an anonymized PDF with identifying metadata removed. Reviewer pages omit author identities and identity disclosures; editorial staff screen files before assignment. Reviewers must not put their names in review comments. Anonymity cannot be guaranteed where subject matter or the file itself reveals identity.
Methods concerns are not outvoted
Acceptance requires all three disciplines. Differing recommendations or major design/statistics concerns trigger independent methods adjudication before acceptance. An editor records the final decision and rationale; the platform does not turn review into an automatic majority vote.
Revisions preserve the record
Authors receive released anonymous review comments with editorial decisions. Revised manuscripts reopen a new review round. Prior uploads, reviews, and decisions remain available to the author or editorial staff as appropriate. Audit events are append-only through the application, not a cryptographically immutable ledger.
Negative results are eligible
Original research, rigorous reviews, methods papers, replications, and negative findings are eligible for editorial consideration. There is no automatic acceptance or rejection based on the direction of findings.
Publication requires context
An accepted paper requires a scholarly text, clinical interpretation, plain-language summary, explicit limitations, and author publication permission. No fabricated studies, impact factors, evidence grades, reviewer counts, or indexing claims appear in the journal.
A permanent place for corrections
Editors append dated corrections, replication links, critiques, evidence updates, and retraction notices to published records. Prior notices are not silently overwritten.
Confidentiality and data handling
Manuscripts and reviewer profiles are stored server-side with role-restricted access. Upload only de-identified research—not patient records or directly identifiable health data. This pilot is not represented as a HIPAA-compliant clinical system. Email notifications are not enabled; check your portal inbox.
Pilot access and governance
This version uses Sign in with ChatGPT and remains private to the owner until access is deliberately expanded. Authors own their drafts. Reviewers require editorial approval. Only authorized editors can assign, decide, publish, and append evidence updates. Clinical interpretations are educational, not treatment recommendations.