Amendments involving changes to IRB-approved protocols do not need prior IRB approval if:

The answer

Changes to an IRB-approved protocol normally require prior IRB approval before they are implemented. There is one recognized exception: an amendment does not need prior approval when the change is necessary to eliminate an apparent immediate hazard to the subject. In an emergency where waiting for IRB review would put a participant's health or well-being at risk, the investigator may make the change right away to protect the subject — and must then report it to the IRB promptly (typically within a few business days, per institutional policy).

This mirrors the federal regulations (45 CFR 46.108 / 21 CFR 56.108), which allow deviations from an approved protocol without prior IRB approval only to eliminate apparent immediate hazards to subjects, followed by prompt reporting.

Why this exception exists

The IRB's core purpose is to protect human subjects. The prior-approval rule normally serves that purpose by vetting changes before they affect participants. But a rigid rule could itself endanger someone if a researcher had to pause and wait for committee review while a subject faced immediate harm. The emergency exception resolves that tension: subject safety comes first, and oversight is preserved through mandatory after-the-fact reporting, so the IRB can review what was done and decide whether the change should continue.

Crucially, this is a one-time, subject-protective measure, not a shortcut for convenience. It doesn't authorize permanent protocol changes; if the emergency change should become standard practice, it must go through the normal amendment process.

Why the other options are wrong

Common distractors describe changes that still require prior approval:

• "The change is minor" — even minor changes generally need IRB review (often via expedited review). "Minor" affects the type of review, not whether approval is required first.
• "The PI thinks it will improve the study" or "it makes enrollment easier" — investigator judgment or administrative convenience never waives prior approval. Only elimination of an immediate hazard to a subject does.
• "It doesn't increase risk" — reduced or unchanged risk still requires prior review; the IRB, not the investigator, makes that determination.

The bigger picture

Distinguish an amendment (a planned, prospective change to the protocol, requiring prior approval) from a deviation (a one-off departure from the approved protocol). The emergency exception is essentially a permitted deviation to protect a subject, reported afterward. Everything else — new procedures, revised consent forms, added study sites, changed eligibility — flows through the standard amendment workflow: submit the change, wait for IRB approval, then implement.