IRB Waiver of Consent: Three Decisions, Five Criteria, and What the 2024 FDA Rule Changed

At a glance

• “A consent waiver” is not one switch. It is three distinct determinations that fail under different rules: which regulatory track governs you, whether you are seeking a full waiver or an alteration, and whether you are waiving consent itself or only the signature that documents it.
• The single biggest change: as of the FDA final rule effective in 2024, an IRB can now waive informed consent for a minimal-risk FDA-regulated investigation under 21 CFR 50.22. Before that, no such pathway existed for FDA studies, so pre-2024 SOPs are wrong in both directions.
• The FDA minimal-risk criteria in 50.22 and the Common Rule criteria in 45 CFR 46.116(f) are now near-identical five-part tests, which is the point of harmonization. State the alignment plainly rather than implying one set governs everything.
• Waiving the signature (21 CFR 50.27 via 56.109(c), or 45 CFR 46.117(c)) is a separate, lower-bar determination from waiving consent. Many teams write “waiver” when they mean a documentation waiver.
• “Not practicable without the waiver” is the criterion IRBs reject on. Cost and convenience are not impracticability.
• This explains what the regulations require. It does not certify any study compliant; the sponsor and investigator remain responsible for the determination they submit.

Three things people call “a consent waiver” (and why mixing them up fails an inspection)

When a coordinator says “we got a consent waiver,” they could mean any of three legally distinct things. The first question is which regulatory track governs the study, because FDA-regulated drug and device investigations run under 21 CFR Parts 50 and 56, while federally conducted or supported research runs under the Common Rule at 45 CFR 46, and the criteria, while now harmonized, are not interchangeable citations. The second question is whether the ask is a full waiver of consent or an alteration that drops or modifies some elements. The third is whether you are waiving consent at all, or only waiving documentation, the signed form. Conflating these is what turns a defensible submission into a finding.

Decision 1: Which regulatory track governs, and what changed in 2024

Start by identifying whether the study is an FDA-regulated clinical investigation. Under FDA rules, the investigator’s baseline duty is firm: 21 CFR 312.60 makes the investigator responsible for obtaining the informed consent of each subject in accordance with Part 50. The general consent requirement in 21 CFR 50.20 states that, except as provided in 50.22, 50.23, and 50.24, no investigator may involve a human subject without legally effective informed consent.

The change practitioners most often miss: 21 CFR 50.22, added by the FDA final rule published December 21, 2023 and carried in the current Part 50 text, now lets the IRB waive or alter consent for a minimal-risk clinical investigation. Before this provision existed, an FDA-regulated study had no minimal-risk waiver pathway at all. The only exceptions were the life-threatening-emergency certification in 50.23 and the planned emergency-research exception in 50.24. That is why stale templates fail in both directions: a pre-2024 SOP that says “FDA studies can never waive consent” is now wrong, and a template copied from the Common Rule that assumed a minimal-risk waiver was always available was, for FDA studies, wrong before 2024.

On the Common Rule side, 45 CFR 46.116(f) has long allowed an IRB to waive or alter consent for minimal-risk research. The two regimes now read almost identically, which is the explicit goal of harmonization. Where both Part 50 and 45 CFR 46 apply to the same study (for example, an NIH-funded, FDA-regulated trial), cite both rather than collapsing them: 50.22 and 46.116(f) impose materially the same five findings, and you should say so plainly. If your IRB applies a stricter institutional SOP, that SOP, not the regulation, is the binding constraint to verify.

Decision 2: Full waiver vs alteration of consent

These are different asks. A full waiver means no consent is obtained from the subject; an alteration means consent is still obtained but some required elements are omitted or modified, for example a deception study that withholds the true purpose until debriefing. Both 21 CFR 50.22 and 45 CFR 46.116(f) cover “waiver or alteration” under the same criteria, so the bar is structurally the same, but the IRB must find and document which one you are requesting. Critically, the Common Rule draws a line you cannot cross: 45 CFR 46.116(f) permits omitting or altering the elements in 46.116(b) and (c), but an IRB may not omit or alter the general requirements in 46.116(a), the bedrock requirements such as no exculpatory language. An alteration is not a license to remove protections; it is a narrowly bounded modification.

Decision 3: Waiver of documentation (signature) vs waiver of consent itself

This is the most commonly mislabeled determination. Waiving documentation means consent still happens, the subject is still informed and still agrees, but the IRB waives the requirement for a signed form. Under FDA rules, 21 CFR 50.27 requires consent to be documented by a written form signed and dated by the subject, except as provided in 21 CFR 56.109(c). And 56.109(c)(1) lets the IRB waive the signature for some or all subjects where the research presents no more than minimal risk and involves no procedures for which written consent is normally required outside the research context. The Common Rule mirrors this: 45 CFR 46.117(c) lets an IRB waive the signed-form requirement on minimal-risk findings, and also where the only record linking subject to research would be the consent form and the principal risk is a confidentiality breach, in which case the subject’s wishes about documentation govern.

The practical takeaway: a documentation waiver is a lower, narrower bar than a consent waiver, and the two are not substitutes. If your subjects will be told everything and will agree but you do not want a signature, you need a documentation waiver, not a consent waiver. Submitting one when you mean the other is a classic finding.

The criteria, line by line

For a minimal-risk consent waiver or alteration, both 21 CFR 50.22 and 45 CFR 46.116(f)(3) require the IRB to find and document five things:

1. Minimal risk. The clinical investigation involves no more than minimal risk to subjects (50.22(a); 46.116(f)(3)(i)). FDA defines minimal risk in 21 CFR 50.20’s definitions as harm or discomfort not greater than that ordinarily encountered in daily life or routine physical or psychological examinations.
2. Not practicable without the waiver. The research could not practicably be carried out without the requested waiver or alteration (50.22(b); 46.116(f)(3)(ii)).
3. Identifiable data/biospecimens. If the study uses identifiable private information or identifiable biospecimens, it could not practicably be carried out without using them in identifiable form (50.22(c); 46.116(f)(3)(iii)). This is the fifth criterion the 2018 Common Rule revision added and that 50.22 now carries.
4. Rights and welfare. The waiver or alteration will not adversely affect the rights and welfare of the subjects (50.22(d); 46.116(f)(3)(iv)).
5. Debrief when appropriate. Whenever appropriate, subjects or their representatives will be provided additional pertinent information after participation (50.22(e); 46.116(f)(3)(v)).

These sit alongside the IRB’s general approval criteria. 21 CFR 56.111 and 45 CFR 46.111 both require, among other findings, that risks are minimized, risks are reasonable relative to benefits, and that consent will be sought and documented or appropriately waived to the extent required by Part 50 or 46.116/46.117 respectively. The waiver does not exempt a study from those baseline approval criteria.

”Not practicable” is the criterion IRBs reject on

Of the five findings, impracticability is where submissions die. The regulations say the research “could not practicably be carried out without” the waiver (50.22(b); 46.116(f)(3)(ii)), and reviewers read that strictly. Cost, slower enrollment, and the inconvenience of re-contacting people are not impracticability. Impracticability usually means the consent process would itself defeat the science (as in deception), or that locating and re-consenting subjects is genuinely infeasible (a retrospective cohort of thousands, many deceased or lost to follow-up, with no current contact path).

A justification that holds reads roughly like this:

This is a retrospective review of 4,200 archived records from encounters between 2009 and 2016. There is no ongoing clinical relationship with these individuals; the institution holds no current contact information for an estimated 78% of the cohort, and prior outreach on an unrelated registry reached under 15%. Re-consent would require successful contact across the full cohort to avoid a selection bias that would invalidate the analysis, and successful contact is not achievable for the majority. The study is therefore not practicable without a waiver. Cost and staff time are not offered as grounds; the obstacle is the absence of a current contact path for most subjects and the bias that partial contact would introduce.

Note what that does: it grounds impracticability in infeasibility and scientific validity, names the magnitude, and explicitly disclaims convenience and cost. That is the shape reviewers accept.

Worked examples and the decision logic

A decision table, in prose, for the common cases:

• Retrospective chart review, minimal risk, FDA-regulated: full consent waiver under 21 CFR 50.22 (post-2024). Identifiable-data criterion (50.22(c)) applies if you need identifiers.
• Retrospective chart review or secondary biospecimen research, minimal risk, Common Rule: full consent waiver under 45 CFR 46.116(f), with the identifiable-data finding under 46.116(f)(3)(iii). Note that if a subject was asked for broad consent and refused, 46.116(f) bars the IRB from then waiving consent for that secondary use.
• Deception study, minimal risk: an alteration, not a full waiver, under 50.22 or 46.116(f), with the post-participation debrief satisfying the “additional pertinent information” element.
• Remote or oral-consent scenario where you will inform and the subject will agree but you do not want a signature: a documentation waiver under 21 CFR 56.109(c) or 45 CFR 46.117(c), not a consent waiver.
• Life-threatening emergency, individual subject, FDA: not a waiver. This is the 21 CFR 50.23 certification, requiring the investigator and an independent physician to certify in writing that the subject faces a life-threatening situation, consent cannot be obtained, time is insufficient to reach a representative, and no alternative therapy offers an equal or greater chance of saving the subject’s life.
• Planned emergency research, no prospect of individual consent at enrollment: the 21 CFR 50.24 exception, a distinct and demanding pathway requiring, among much else, that the research could not practicably be carried out without the waiver (50.24(a)(4)) plus community consultation and public disclosure. It is not a minimal-risk waiver and must run under a separate IND or IDE per 50.24(d).

ICH E6(R3) sits over all of this as the GCP overlay. E6(R3) section 2.1 requires that freely given informed consent be obtained and documented from every participant prior to participation, with the legally acceptable representative consenting where the participant cannot. E6(R3) section 2.8.1 requires the investigator to comply with applicable regulatory requirements and to adhere to GCP and the ethical principles of the Declaration of Helsinki when obtaining and documenting consent. For emergency situations where prior consent is not possible, E6(R3) section 2.8.8 requires enrollment to follow measures described in the protocol with documented IRB/IEC approval, with the participant or representative informed as soon as possible. There is a tension worth stating plainly rather than smoothing over: E6(R3) frames consent as something to be obtained and documented from every participant and is built around emergency and representative-consent scenarios, whereas the FDA and Common Rule minimal-risk waiver provisions (50.22; 46.116(f)) authorize dispensing with consent entirely for qualifying low-risk research. The GCP guideline does not itself grant a minimal-risk waiver; the statutory waiver authority lives in Parts 50/56 and 45 CFR 46, and the investigator must satisfy whichever applies while still meeting GCP. Cite both and let the reviewer see the boundary.

What the IRB submission must contain, and the common findings

Your submission should make the IRB’s job mechanical: state which track applies, state whether you seek a waiver, an alteration, or only a documentation waiver, and then walk the five findings (50.22 or 46.116(f)(3)) one by one with the impracticability paragraph written to the standard above. Where both FDA and Common Rule apply, address both citations.

Where teams get it wrong, and where the findings cluster:

• Labeling a documentation waiver as a consent waiver, or vice versa. These are different determinations under different sections (50.27/56.109(c) and 46.117(c) for documentation; 50.22 and 46.116(f) for consent). Cite the one you actually need.
• Resting impracticability on cost or convenience. The text demands that the study “could not practicably be carried out without” the waiver; reviewers reject convenience.
• Relying on a stale SOP that misstates FDA authority, either denying that 50.22 exists or assuming a minimal-risk waiver was always available for FDA studies.
• Treating an alteration as license to strip protections. Under 45 CFR 46.116(f), the general requirements in 46.116(a) may never be omitted or altered.
• Forgetting that, separate from any waiver, the investigator must follow the IRB-approved process and may not change the research without IRB approval. 21 CFR 312.66 requires the investigator to assure IRB review and to make no changes in the research without IRB approval except to eliminate immediate hazards, and ICH E6(R3) section 2.5.4 likewise allows deviation from the protocol only where necessary to eliminate an immediate hazard to participants.

A consent waiver, properly determined, is a legitimate and common tool. Getting it inspected cleanly is a matter of naming the right one of three decisions, satisfying the five findings on the record, and writing impracticability so it withstands scrutiny. Verify the specifics against your IRB’s current SOP before you submit.

Sources

• ICH E6(R3) Good Clinical Practice, version r3 — ICH — https://www.ich.org/page/efficacy-guidelines
• 21 CFR Part 50 Protection of Human Subjects, version 2024 — FDA — https://www.govinfo.gov/content/pkg/CFR-2024-title21-vol1/pdf/CFR-2024-title21-vol1-part50.pdf
• 21 CFR Part 56 Institutional Review Boards, version 2024 — FDA — https://www.govinfo.gov/content/pkg/CFR-2024-title21-vol1/pdf/CFR-2024-title21-vol1-part56.pdf
• 45 CFR 46 Protection of Human Subjects (Common Rule), version 2018 — HHS — https://www.govinfo.gov/content/pkg/CFR-2024-title45-vol1/pdf/CFR-2024-title45-vol1-part46.pdf
• 21 CFR Part 312 Investigational New Drug Application, version 2026-04 — FDA