Clinical safety as a publication standard: review gates tied to on-chain execution

The history of medicine offers few unalloyed lessons. One of them is this: instruments that record what happened are more durable than instruments that record what was intended. This paper concerns the procedural geometry that holds intentions and outcomes in the same file.

A protocol that disburses funds to interventions touching patient populations cannot relax the standards by which clinical safety is assessed. Nor can it relax the conditions under which an assessment is considered to have been performed. We propose a publication-grade treatment of clinical safety review, in which the act of review is itself a publishable instrument with the same standing as the underlying proposal.

I. Premises

Three premises orient the standard. First, clinical safety is not a checklist; it is a judgment exercised under named conditions by named reviewers. Second, the conditions under which the judgment was exercised — disclosures, references, points of dissent — must be recorded with the judgment itself. Third, the record must persist for as long as the intervention persists in any descendant data.

II. The instrument

The instrument is composed of four parts: a written review, an appendix of disclosures, a register of dissents, and a post-execution report keyed to the execution record. Each part is signed by the reviewer or reviewers responsible. The instrument is filed against the proposal’s primary key in the protocol registry.

The procedural distance between “the trial was safe” and “the trial was conducted under conditions in which the question of safety could be meaningfully assessed” is the distance this instrument is intended to traverse.

III. Execution gates

The on-chain execution of any proposal that touches patient populations is gated on the presence of a filed instrument that conforms to the standard. The gate is not a votable parameter; it is a property of the registry. Proposals that pass deliberation but fail the gate are recorded as such and revert to second reading with the deficiency named.

IV. Failure modes

We anticipate three failure modes: a reviewer signs in form but not in substance; the appendix of disclosures is incomplete because the reviewer did not know what they did not know; a register of dissents is suppressed by social pressure within a small reviewing pool. Each mode is addressed in §VI of the standard.

V. On the question of speed

A standard of this density slows the protocol. We consider this a feature. Capital that touches bodies should move with friction; the friction is the function. Where speed is required — emergency standby, defensive funding — a separate procedure applies, scoped narrowly and reviewed quarterly.