Regulation — pharma-specific
IRB Common Rule (45 CFR 46) — continuing review of AI-driven eligibility
The Common Rule places continuing-review duty on the IRB. When AI participates in eligibility, the IRB's questions about equitable subject selection and risk-benefit balance land on the sponsor. A silent exclusion pattern is the exact input that triggers IRB continuing-review escalation.
What the regulation actually says
"In order to approve research covered by this policy the IRB shall determine that … selection of subjects is equitable … and appropriate additional safeguards have been included in the study to protect the rights and welfare of these subjects."
45 CFR §46.111(a)(3) — equitable-selection criterion
"An IRB shall conduct continuing review of research requiring review by the convened IRB at intervals appropriate to the degree of risk … not less than once per year."
45 CFR §46.109(e) — continuing-review duty
What this means in plain English
Two Common Rule criteria that AI-driven eligibility directly implicates:
- Equitable selection. If AI screens candidates and systematically excludes an underrepresented subgroup, the equitable-selection finding on the original protocol may no longer hold.
- Continuing review. The IRB's continuing review will ask "has anything changed?" A silent AI-driven exclusion pattern that emerged post-approval is exactly the change that requires disclosure.
Sponsors that discover the pattern via their own audit have a very different disclosure conversation than sponsors that discover it via IRB question.
What triggers the exposure in the sample
P3 exclude-lane rate rose from 22% to 75% silently over the audit window. That pattern reaches the equitable-selection criterion directly. An IRB continuing-review board that learns of this via sponsor disclosure hears a very different case than a board that learns of it via participant complaint or auditor inquiry — and trial-site suspension is a live consequence in the second scenario.
What the $499 Snapshot shows against this rule
- Dated evidence of when the drift emerged — the "sponsor learned on Day X" record
- Per-patient-group exclusion quantification — direct input to equitable-selection re-review
- 3 fix-first items scoped to sponsor operational response — the additional-safeguards restart
- Signed independent-verifier declaration — supports the sponsor-initiated-disclosure stance
See the exclusion-differential chart →
$499 Snapshot. 3 business days.
Independent-verifier determination scoped to your sponsor's AI surface + 3 fix-first items + signed declaration.
Buy $499
Snapshot credit applies to Baseline ($2,500) or Enterprise Attestation ($35-55K) upgrade within 30 days.