For journals and editorial offices
Every submission asks for a disclosure statement, and every author writes one out again from recall. Nobody checks it, because until now there was nothing to check it against.
PayClear reconciles what an author submitted against what is on the record, per author, and returns an answer an editor can act on. The manuscript itself never leaves your system: we receive the author identifiers and the disclosure statement being checked, and nothing else. We are not a plagiarism check and not an integrity screen. One narrow question, answered well.
Every other record in this market is a company checking its own homework. Ours is the only one that asks the person whose name is on it.
A radiation oncologist we work with keeps a careful disclosure record. His filled ICMJE form lists an honorarium from a company. The prose statement he publishes on papers does not.
Two documents, one relationship, quietly disagreeing. He is diligent. He maintains the same disclosure by hand in several places over years, which is exactly how this happens, and it is invisible to an editor reading either one on its own.
Four authors, four disclosure statements, one answer each.
Illustrative example with fictional authors and fictional results. Nothing here is a real physician, a real ledger, or a real journal.
One page per manuscript, reached from a link beside the submission. No account, no login, nothing to install. Your editors have never heard of us and never need to.
Your submission system already carries automated checks as links in a manuscript's action list, each showing its answer next to the link so an editor can triage a queue without opening anything. This is built to sit in that same list, with its count beside it. We are not a similarity check and do not do its job. We borrowed its shape, because editors already know how to read it.
The link reads 2 of 4 need review. It will never read as a percentage. A single number against a named person becomes a verdict on that person the moment anyone glances at it, and what we produce is a difference to resolve, not a score to act on.
We hold your reference for it and the author identifiers. Not the title, not the abstract, not a line of the text. A shorter security review is the point.
Each report opens on a single-purpose link tied to one manuscript, revocable on its own without disturbing anything else. Treat it as confidential, the way you would the rest of the submission file.
How each author's submitted statement reconciles against the record. Differences are for you to raise with the author. They are not findings of misconduct.
A difference on an author we are certain we identified. Reconcile it with them.
2 entities on record and absent from the submitted statement, totalling $41,800.00.
Something did not reconcile, or the identity is not certain. A question for the author.
Possible match, identity not confirmed. Specific companies and amounts are withheld until this author is confirmed, because they may belong to a different person with the same name.
Based on US public payment data. Payments from manufacturers outside the ingested corpus, non-US payments, and categories the federal record does not collect are not covered by this answer.
Authors appear as your own reference for them, not by name. Your system already knows who Author 3 is, and a page about someone's financial relationships is safer without their name printed across it.
Illustrative. The journal, manuscript, authors and companies are invented, and the figures with them. A real report is reached by a link issued to your submission system and is not publicly addressable.
A binary pass or fail forces a machine to make a call it cannot support. The middle tier is where uncertainty goes, and most non-clear answers belong there.
Everything on the author's record appears in the statement they submitted. The answer also states how much we could see, so a thin result cannot read as a clean one.
Something did not reconcile, or the person could not be identified with certainty. This tier exists to absorb doubt. Most answers that are not cleared land here, and that is by design.
A real gap on an author we are certain we identified. Never reached on a name match alone, however unusual the name.
These matter more than the features. A screening tool that raises a conflict about the wrong person gets switched off, and it only has to happen once or twice.
Two physicians share a name far more often than they share a name and an institution. Until identity is confirmed, the specific companies and the amount are withheld and you are told why. You still get the count and the instruction to confirm, which is everything you need to ask the author. Printing a stranger's payers under your author's name is not a smaller mistake than missing one.
The word is unreconciled. A difference between a record and a statement is a question for the author, not a finding against them. An editor gets a reconciliation difference to raise, never an accusation.
Nothing reaches the top tier unless identity is certain: an author-issued token or an exact NPI. A unique name match is good evidence and still not proof, so it caps at needs review. This is the rule that keeps the integration installed.
A reviewer never chose to be screened, so reviewer checking is off unless the person turns it on. When it does run it returns a verdict and a reason code and no records at all, so results cannot be accumulated into a list of conflicted reviewers.
Without a token from the author, a check sees public record data and, at most, whether a ledger exists. The deepest answers are the ones the author chose to give you.
The federal Open Payments record for the three program years ICMJE asks about, compiled and kept locally so a check makes no outside call while your editor waits.
One row per physician, per paying company, per program year.
Everyone CMS names as a covered recipient in the window, not only those with a PayClear account.
Every manufacturer and GPO in the record. A company needs no relationship with us to appear.
The three published years, matching the 36 months ICMJE asks about.
General payments only, so research payments and ownership interests are not in these figures. US federal reporting only, so a non-US author, a non-prescriber, or anyone who has received nothing reportable will come back as no match. That is not a clean result and we never render it as one. Every answer carries the reach of the search alongside it.
A check reads a record we have already compiled and makes no outside calls while your editor waits. An unknown author comes back as still compiling and is delivered later, rather than holding up a decision.
AAMC Convey holds what an author has entered. PayClear reconciles that against what is on the record elsewhere. Different questions, and we would rather sit alongside it than ask anyone to replace it.
An author we cannot resolve today is compiled afterwards, so the next journal asking about that person gets a better answer. Nothing about a manuscript is retained to do it.
The rail is built and running. We are looking for a small number of editorial offices to shape it before it is priced, which means the first partners get to argue with the defaults while they are still cheap to change. We would rather say that than imply a queue.
Worth a conversation? michael@payclear.health