Central IRB review uses one independent institutional review board to approve a multi-site protocol for every participating site, while local IRB review requires each institution to conduct its own separate approval. Under the NIH single-IRB (sIRB) policy and the revised Common Rule, most US federally funded multi-site human subjects research must now use a single IRB of record rather than parallel local reviews, with reliance agreements defining how participating sites delegate oversight.
An Institutional Review Board (IRB) is a federally mandated committee, operating under 45 CFR 46 and 21 CFR 56, that reviews research involving human subjects to protect participants’ rights and welfare. The central-versus-local choice is not cosmetic: it changes who signs off on the protocol, how long approval takes, and what your research administration office must manage.
Contents
- What is the difference between a central IRB and a local IRB?
- When does NIH require a single IRB for multi-site research?
- How do IRB reliance agreements work?
- Central vs local IRB: practical trade-offs for research administrators
- IRB vs REB vs REC: terminology across the US, Canada, and the UK
- Frequently asked questions
- Implications for research administrators
What is the difference between a central IRB and a local IRB?
A local IRB is a review board affiliated with a single institution — a university, hospital, or academic medical centre — that reviews research conducted at that institution alone. A central IRB (sometimes called a commercial or independent IRB) is an external board that provides review services to multiple institutions participating in the same multi-site protocol, issuing one approval that all relying sites can adopt.
Local IRBs bring institution-specific knowledge: familiarity with the local patient population, policy, and community standards. Central IRBs bring standardisation: one protocol interpretation, one informed-consent template, one review timeline across every site. Both apply the same federal review criteria and carry equal regulatory authority — the difference is organisational scope, not rigour:
- Local IRB — institution-specific; retains ancillary reviews (biosafety, conflict of interest, contract language)
- Central IRB — multi-institutional; reviews on behalf of all participating sites under one protocol
When does NIH require a single IRB for multi-site research?
The NIH single-IRB (sIRB) policy applies to NIH-funded, non-exempt, multi-site human subjects research at more than one domestic site, and took effect for grant applications submitted on or after 25 January 2018. One IRB of record — a central IRB or a designated local IRB at one participating site — must review the protocol for every US site.
The revised Common Rule (the “2018 Requirements” to 45 CFR 46) extended a parallel single-IRB requirement to most federally funded cooperative multi-site research, with compliance required from 20 January 2020. Exceptions are narrow: NIH requires a compelling justification, such as conflicting state law, and explicitly states that cost alone is not an acceptable reason for an exception.
A related but distinct point: the FDA does not mandate single IRB review under its own regulations at 21 CFR 56. FDA guidance encourages centralised review for efficiency in FDA-regulated multicentre trials, but the binding requirement comes from NIH grant policy and the Common Rule, not FDA rulemaking — a distinction administrators often need to clarify for investigators.
How do IRB reliance agreements work?
An IRB reliance agreement is a formal document — typically an IRB Authorization Agreement (IAA) — in which a participating site defers its own ethical review to the designated IRB of record. The relying site keeps responsibility for local context review: conflicts of interest, HIPAA compliance, and institutional biosafety, outside the sIRB’s regulatory scope.
SMART IRB, a national reliance platform used by hundreds of US institutions, standardises this process with a single master agreement signed once, then invoked per-study rather than renegotiated for every protocol. This is the single biggest driver of variance in real-world turnaround time: pre-signed institutions can activate a relying-site study in days, while a one-off IAA negotiation can add weeks regardless of how fast the central IRB itself reviews the protocol.
Central vs local IRB: practical trade-offs for research administrators
For research administrators, the practical difference rarely comes down to review quality — both apply identical federal criteria — and comes down instead to coordination overhead. A peer-reviewed analysis published via PubMed found central IRB review shortened submission-to-approval time versus local review, though total time to study activation still depended on how quickly each site’s reliance agreement was completed.
| Factor | Local IRB | Central IRB |
|---|---|---|
| Review consistency across sites | Variable — each site interprets independently | Standardised — one interpretation for all sites |
| Typical approval-stage speed | Slower for multi-site studies; duplicated review effort | Faster approval stage; single review cycle |
| Local context knowledge | Strong — direct institutional familiarity | Delegated back to relying site via local context review |
| Cost structure | Often bundled into institutional overhead | Often a separate per-study or per-site fee |
| Best suited to | Single-site studies; institution-specific research | Federally funded multi-site trials subject to sIRB mandate |
Institutions that under-resource the reliance-agreement function — treating it as an afterthought rather than a parallel workstream — are the most common source of activation delay under the sIRB model, even when the central IRB itself performs efficiently.
IRB vs REB vs REC: terminology across the US, Canada, and the UK
Research administrators working across borders must distinguish three related but jurisdictionally distinct terms. In the United States, the governing term is IRB (Institutional Review Board), defined under 45 CFR 46 and 21 CFR 56. In Canada, the equivalent is a REB (Research Ethics Board), operating under the Tri-Council Policy Statement (TCPS 2) issued jointly by CIHR, NSERC, and SSHRC. In the United Kingdom, the equivalent is a REC (Research Ethics Committee), coordinated through the Health Research Authority (HRA) for NHS-related research.
The three terms are functionally analogous — each reviews human subjects research for ethical acceptability — but they sit inside different statutory frameworks and are not interchangeable for compliance purposes. A protocol approved by a US central IRB does not satisfy a UK REC’s approval requirement, and vice versa; multinational studies typically need separate approvals under each jurisdiction’s framework, connected where possible through mutual recognition arrangements rather than a single cross-border reliance agreement.
Frequently asked questions
What is the difference between central IRB and local IRB?
A central IRB reviews a protocol once on behalf of every site in a multi-site study, while a local IRB reviews only research conducted at its own institution. Both apply identical federal review standards; the difference is organisational reach, not the rigour of the ethical review itself.
What are the three types of IRB review?
US regulations define three review paths: exempt (minimal-risk categories defined by regulation), expedited (minimal-risk research reviewed by one or more designated reviewers rather than the full board), and full board review for studies presenting greater than minimal risk. The path is determined by risk level, not by whether the IRB is central or local.
What is the difference between IRB of record and central IRB?
The IRB of record is whichever board holds review authority for a given study — a central/commercial IRB, or a participating institution’s own local IRB acting in that role. “Central IRB” describes an operating model; “IRB of record” describes a regulatory designation a central IRB usually, but not always, fills.
What is the difference between local IRB and external IRB?
An external IRB is any IRB outside a site’s own institution designated to hold review authority under a reliance agreement — frequently a central IRB, but sometimes another institution’s local IRB. The relying site’s own human research protection office typically still reviews matters such as conflicts of interest and HIPAA compliance.
Implications for research administrators
As the sIRB mandate matures, the burden on research administration offices has shifted from running IRB review itself toward managing reliance infrastructure: pre-negotiating master agreements, tracking which sites have signed onto platforms such as SMART IRB, and running local context review in parallel with central IRB approval rather than after it.
Institutions that treat reliance-agreement readiness as a standing capability, not a one-off task, activate multi-site protocols faster than those negotiating terms study by study. The single-IRB requirement is not going away; institutions that adapt their reliance workflow now will realise the turnaround-time benefit the policy was designed to deliver.
For definitions of related human-subjects and research-administration terminology, see the CASRAI Dictionary. For broader context on research administration standards and workflow interoperability, see CASRAI’s research administration content hub.