IRB (Institutional Review Board)

An IRB (Institutional Review Board) is the US statutory body responsible for prospective and continuing review of human-subjects research, established to ensure that risks to participants are minimized and reasonable relative to anticipated benefits. Investigators may not enroll participants, collect identifiable data, or use identifiable biospecimens until they hold an unexpired IRB approval letter covering the protocol version they intend to execute. Continuing review, amendment review, and reportable-event review remain ongoing obligations for the life of the study.

Legal Basis: The Common Rule and FDA Regulations

The IRB requirement rests on two parallel federal regulatory tracks that frequently apply to the same study:

45 CFR 46 (the Common Rule) — the general federal policy for human-subjects research conducted or funded by any of the federal departments/agencies that have adopted it. Subpart A is the core policy; Subparts B, C, and D add protections for pregnant women/fetuses/neonates, prisoners, and children respectively (see the site’s dedicated 45 CFR 46 Subpart D term for the pediatric research track).
21 CFR Part 56 — FDA’s parallel IRB regulation, which governs IRB review of clinical investigations of drugs, biologics, and devices regulated by FDA, regardless of funding source. A study can be subject to 21 CFR 56 alone, 45 CFR 46 alone, or both at once (e.g., an NIH-funded drug trial), in which case both regulatory tracks’ requirements apply independently.

Institutions conducting HHS-supported human-subjects research must hold a Federalwide Assurance (FWA) with the Office for Human Research Protections (OHRP), and the reviewing IRB itself must be separately registered with OHRP before it has authority to approve covered research — registration and the FWA are two distinct, sequential filings.

The Three Types of IRB Review

Not every human-subjects protocol goes to a convened board meeting. The Common Rule sets up three review pathways, and which one applies is a determination the IRB (not the investigator) makes based on the nature of the research and its risk level:

Exempt. Research that fits one of the categories set out at 45 CFR 46.104 (for example, certain research on normal educational practices, some anonymous survey/interview research, or secondary use of existing identifiable data under specified conditions) is exempt from the Common Rule’s general review requirements — but the exemption determination itself is still made by the IRB or another designated institutional official, not self-certified by the investigator.
Expedited. Research that is no more than minimal risk and fits one of OHRP’s enumerated expedited-review categories (minor blood/specimen collection, data via noninvasive procedures routine in clinical practice, certain survey/interview research, minor changes to previously approved research, and similar low-risk activities) can be reviewed by the IRB Chair or a Chair-designated experienced reviewer under 45 CFR 46.110 / 21 CFR 56.110, without waiting for a convened meeting. An expedited reviewer may exercise all the authorities of the full board except disapproval — a study can only be disapproved through full-board review at a convened meeting. See the dedicated expedited review term and the expedited vs. full board comparison for the category list and typical turnaround differences.
Full board (convened) review. Research that is more than minimal risk, or that does not fit an expedited category, must be reviewed at a convened meeting under 45 CFR 46.108, with a majority of members present, including at least one non-scientist member, and approved by a majority vote of those present.

IRB Membership and Composition Requirements

45 CFR 46.107 sets minimum composition rules designed to ensure the board can competently evaluate both the science and the ethics of a protocol:

At least five members, with varying backgrounds, sufficiently qualified through the collective experience, expertise, and diversity of the members to review the research the institution commonly conducts.
At least one member whose primary concerns are scientific and at least one member whose primary concerns are nonscientific (e.g., an ethicist, lawyer, or clergy member).
At least one member not otherwise affiliated with the institution and not part of the immediate family of anyone who is affiliated with it — the “non-affiliated member” requirement, meant to bring a genuinely outside perspective to the board.
Where an IRB regularly reviews research involving populations vulnerable to coercion or undue influence (children, prisoners, individuals with impaired decision-making capacity, or economically/educationally disadvantaged persons), consideration must be given to including members knowledgeable about and experienced in working with those populations. Reviewing prisoner research specifically requires a prisoner representative on the board under 45 CFR 46 Subpart C.
A member with a conflicting interest in a specific protocol may supply information the board requests but may not otherwise participate in review or vote on that protocol (46.107(d)).
The IRB may invite outside individuals with specialized expertise to assist review, but they may not vote.

See the site’s IRB membership requirements guide for a fuller walkthrough, and the IRB Chair term for how the chair role and expedited-review authority fit into this structure.

What the IRB Actually Evaluates: The Approval Criteria

45 CFR 46.111 sets out the substantive criteria an IRB must find satisfied before it can approve research. In summary, the IRB must determine that:

Risks to subjects are minimized by using procedures consistent with sound research design that do not unnecessarily expose subjects to risk, including using procedures already being performed on subjects for diagnostic or treatment purposes where appropriate (46.111(a)(1)).
Risks to subjects are reasonable in relation to anticipated benefits, if any, and the importance of the knowledge that may reasonably be expected to result (46.111(a)(2)) — the core risk-benefit determination.
Selection of subjects is equitable.
Informed consent will be sought from each prospective subject or their legally authorized representative, and appropriately documented, in accordance with 45 CFR 46.116/46.117 (see the site’s informed consent term and its related waiver of consent guide for the consent-specific requirements).
Where appropriate, the research plan makes adequate provision for monitoring the data collected to ensure subject safety.
Where appropriate, there are adequate provisions to protect the privacy of subjects and maintain the confidentiality of data.
When some or all subjects are likely to be vulnerable to coercion or undue influence, additional safeguards are included to protect their rights and welfare (46.111(b)) — the general criterion that underlies the more specific Subpart B/C/D protections for pregnant women, prisoners, and children.

Continuing Review

45 CFR 46.109(e) requires an IRB to conduct continuing review of research requiring convened-IRB review at intervals appropriate to the degree of risk, and not less than once per year, unless one of the exceptions below applies. The 2018 revision of the Common Rule narrowed when continuing review is actually required: under 46.109(f), continuing review is not required (unless the IRB determines otherwise) for research that is eligible for expedited review, research reviewed under certain limited-IRB-review exempt categories, or research that has progressed to the point where it involves only data analysis (including of identifiable private information or biospecimens) or accessing follow-up clinical data from procedures subjects would undergo as part of clinical care. Before the 2018 revision, continuing review was the default expectation across a much broader set of ongoing studies, including many that had already moved into a data-analysis-only phase. See the site’s dedicated continuing review term and continuing review guide for the submission mechanics and lapse-in-approval consequences.

Single IRB and Reliance Arrangements

For multi-site studies, federal policy increasingly favors designating one IRB of record — a single IRB (sIRB) — rather than requiring separate local IRB review at every participating site. The NIH single-IRB policy (effective January 2018) requires NIH-funded multi-site studies conducting the same protocol to rely on one IRB of record, subject to limited exceptions (e.g., a local law requirement, or the funding component determining sIRB review is not appropriate); the 2018 Common Rule revision added a parallel general expectation for cooperative research. Sites still typically retain responsibility for local-context review (e.g., local law, local investigator qualifications) and for the consent process itself, even once review authority is centralized. See the site’s Central IRB / Single IRB (sIRB) term and the single IRB vs. local IRB comparison for how reliance agreements (including the IRB Authorization Agreement / IAA mechanism) are structured.

IRB vs. IEC vs. REB: Terminology Across Jurisdictions

“IRB” is the US term. Outside the United States, the same functional role — prospective ethics/scientific review of human-subjects research — goes by different names tied to different national regulatory frameworks: REC (Research Ethics Committee, UK, under the NHS Health Research Authority), REB (Research Ethics Board, Canada, under the Tri-Council Policy Statement, TCPS2), HREC (Human Research Ethics Committee, Australia, under the NHMRC National Statement), and IEC (Independent Ethics Committee), a term used broadly in ICH E6 Good Clinical Practice guidance and in many non-US regulatory contexts for the ethics-committee role in a clinical trial. None of these are identical in their governing regulations, but they occupy the same structural position in the research-oversight process. See the site’s IRB vs. Ethics Committee comparison for the fuller jurisdiction-by-jurisdiction breakdown.

What an Investigator Actually Submits

A typical initial IRB submission includes the research protocol, the informed consent document(s) (or a request to waive consent or its documentation), any recruitment materials (flyers, scripts, advertisements), data collection instruments, and information about the investigator’s qualifications and any relevant conflicts of interest. Once a study is approved, investigators have ongoing obligations to submit continuing review materials (where required), protocol amendments for IRB approval before implementation, and reportable-event notifications (e.g., unanticipated problems involving risks to subjects). See the site’s IRB/REC approval process guide for a step-by-step walkthrough of what gets reviewed and what doesn’t.

Related CASRAI Resources

Primary Sources

Common Rule, 45 CFR 46 (2018 revision), Subpart A §46.107 IRB membership, §46.108 IRB functions/meetings, §46.109 IRB review of research, §46.111 criteria for IRB approval
FDA, 21 CFR Part 56, Institutional Review Boards
HHS Office for Human Research Protections (OHRP) IRB Guidebook and Federalwide Assurance/IRB Registration guidance

Every approved study is assigned an IRB number, which identifies the protocol in all later correspondence.