Standing Evidence Panel

Charter summary

The Standing Evidence Panel independently reviews GRADE certainty and evidence anchors on ARKA's clinical ratings. No rating is promoted on a single-clinician approval. The panel adopts the IOM 2011 conflict standard by name: prior disclosure, divestment where possible, a minority of the panel conflicted, and a chair and co-chair entirely free of conflict.

Charter version 1.0.0. The full charter — purpose, composition, COI standard, disclosure, recusal, scope, quorum, publication, and term — is rendered in the sections below and maintained as the versioned source at docs/governance/EVIDENCE_PANEL_CHARTER.md.

Roster

Each person who influences an ARKA rating is named here with role, affiliation, seat, and term. Vacant seats render empty — never a placeholder name. Source: docs/governance/panel-roster.json.

Chair conflict status

Chair seat: vacant. IOM 2011 requires the chair and any co-chair to be entirely free of conflict; a conflicted chair cannot sit and the session cannot proceed. When a chair is appointed, an empty disclosure is published as none declared, never omitted.

Conflict register

Declared interests and assigned COI class for each seated member, published on this page — not behind a link or a modal. An empty disclosure is rendered as none declared rather than omitted. Dollar amounts and private contact details are withheld.

No members are seated yet. Required seats appear empty on the roster above. When members are appointed, each row below will show their interests or none declared.

Session schedule

Four ninety-minute sessions cover the top fifty queue rows. Published panel records are linked below when they exist.

Composition and seats

Four required seats: radiology, primary care, an evidence methodologist with GRADE expertise, and a patient representative. No more than half of voting seats may be held by radiology specialists. No person with an equity interest in ARKA may hold a seat.

Members are appointed for renewable two-year terms after prior disclosure and divestment where possible. ARKA employment is a conflict: an employee may never chair or co-chair.

COI standard (IOM 2011)

The panel adopts the Institute of Medicine's 2011 Clinical Practice Guidelines We Can Trust standard, restated as four binding rules: prior disclosure before appointment; divestment of conflicting financial interests where possible; a minority of the panel holding any remaining conflict; and a chair and co-chair entirely free of conflict.

Each item requires a fresh disclosure, including an explicit “None.” A relevant conflict requires recusal; recusals are published on the panel record by role and conflict class. An unrecused relevant conflict invalidates sign-off.

Scope

The panel adjudicates GRADE certainty and evidence anchors on ARKA's clinical ratings. It does not set utilisation targets, does not review coverage criteria, does not see any customer data, and does not advise the company on commercial matters.

Decision procedure and quorum

Sign-off requires approvals from at least three distinct roles, including the evidence methodologist and at least one non-radiology role, with an unconflicted chair present. Conflicted members must be a minority of those seated. Abstentions and recusals do not count.

ARKA may not reverse a panel decision silently. An override is published with its reason. An unpublished override is void.

Publication

Panel decisions are published, including adverse decisions. The authoritative artefact is the versioned, hashed panel record, which pins the charter version in force on the session date. A human-readable entry is also appended to docs/CLINICAL_SIGN_OFF_LOG.md.

Review queue (top 50)

The panel cannot review all 1,536 ratings in one sitting. The queue is ordered by expected harm — order volume × certainty deficit × recommendation strength — computed from published low-value measure denominators and the DE-SynPUF fixture, never hand-curated. Rankings ARKA is already confident about are not dropped: over-confident high-volume ratings remain in the function. The top fifty fill four ninety-minute sessions. The full method and top 200 rows are in docs/governance/review-queue.json.

Published records

• epr-2026-08-23-cardiac-preop-arrival · session 2026-08-23 · charter 1.0.0 · 6 decisions · hash sha256:bbcd860ddbdf00f202db9fcf8df4e4b915caaa8cc3bb1eed716efefb7970275f