The 2018 Common Rule delay and NIH’s single IRB start left research offices running four clocks at once
On January 22, 2018, federal agencies published a six-month delay of the revised Common Rule, past its scheduled January 19 start. On January 25, NIH’s single IRB policy took effect for grant applications. In our reading, research offices ran four clocks at once.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
On January 22, 2018, federal agencies published a six-month delay of the revised Common Rule, past its scheduled January 19 start. On January 25, NIH’s single IRB policy took effect for grant applications. In our reading, research offices ran four clocks at once.
- The week On Monday, January 22, 2018, the Federal Register published an interim final rule delaying the revised Common Rule.
- The rulebook clock: which Common Rule governed a study The first clock decided which rules applied.
- The review clock: single IRB under two different rules The second clock governed who reviewed multi-site research.
- The systems clock: forms, templates and software for two rulebooks The third clock governed the institution’s tools.
- The money clock: who paid for the change The fourth clock concerned cost.
- How this differed across institutions The four clocks ran in every institution subject to the Common Rule.
- Objections, in order We take five objections, numbered in the order we address them.
- What a research leader could check These checks follow from the four clocks.
- With hindsight The further delay came.
The week
On Monday, January 22, 2018, the Federal Register published an interim final rule delaying the revised Common Rule. Its title described a delay of the revisions to the Federal Policy for the Protection of Human Subjects. It was issued jointly by the Department of Health and Human Services and other Common Rule departments and agencies. It delayed the revised Common Rule, which the rule called the “2018 Requirements,” by six months.
The revisions dated from a final rule published on January 19, 2017. That rule set an effective date and general compliance date of January 19, 2018. It set a separate compliance date of January 20, 2020, for the cooperative research provision, the requirement that U.S. institutions in multi-site research rely on a single institutional review board. So the January 22 publication came after the date on which the revisions had been scheduled to take effect.
The interim final rule explained why. After the 2017 rule was published, it said, representatives of the regulated community had expressed concern about their ability to implement all of the 2018 Requirements by the scheduled date. Some asked for a delay, except for certain burden-reducing provisions. The rule named those provisions: carve-outs from the definition of research, exemptions, elimination of continuing review for certain research, and elimination of the requirement that IRBs review grant applications. It said HHS’s Secretary’s Advisory Committee on Human Research Protections had recommended a delay in August 2017.
The rule moved the effective date and general compliance date to July 19, 2018. Until then, it said, regulated entities would continue to comply with the pre-2018 Requirements, and those requirements would be enforced. It added that regulated entities were not allowed, before July 19, 2018, to comply with the 2018 Requirements in place of the pre-2018 Requirements. Studies started before July 19 would by default stay under the pre-2018 Requirements for their duration. After that date, institutions could elect to move them to the new rules.
The rule did not delay the cooperative research provision. Its compliance date remained January 20, 2020.
The agencies invoked the good cause exception to notice and comment. They said the regulated community and HHS’s advisory committee had cited the rule’s complexity, the absence of needed guidance, and the need to revamp institutional procedures and electronic systems, and that they agreed more time was needed. They said institutions that hastily implemented the rule without preparation were bound to make mistakes that could jeopardize research and the safety of subjects. They also said they were developing a proposed rule on a further delay, to January 21, 2019. Comments on the interim rule were due March 19, 2018.
The rule’s regulatory analysis estimated the effect of the six-month delay. Its summary table, in millions of 2016 dollars annualized, showed quantified benefits, meaning cost savings, of 7.4 at a 3 percent discount rate. It showed quantified costs, meaning foregone benefits, of 49.5. Foregone benefits were concentrated in categories such as the expansion of exempt research, the elimination of IRB review of grant applications and the elimination of some continuing review.
On Thursday, January 25, a separate federal policy took effect. The National Institutes of Health’s policy on the use of a single IRB for multi-site research, issued in June 2016, had been scheduled to start on May 25, 2017. NIH had moved the start to January 25, 2018. Its June 2017 notice said the policy would apply to all competing grant applications with due dates on or after January 25, 2018, and to R&D contract solicitations issued on or after that date.
The NIH policy set an expectation that domestic sites in NIH-funded, non-exempt, multi-site studies using the same protocol would rely on a single IRB of record. Applicants were to submit a plan naming the single IRB, and NIH’s acceptance of the plan would become a term of the award. The policy did not apply to career development, research training or fellowship awards, or to foreign sites.
A delay sounds like nothing happening. For research institutions, the week of January 22, 2018, was the opposite. One federal rule was postponed after its own start date had passed. A related federal funding policy began on January 25. A third date, for the Common Rule’s own single IRB mandate, stayed where it was. Research offices had to track several dates at once, each governing a different part of their work.
In our view, the week showed that a research compliance program is a set of clocks, not a single rulebook. A study’s governing rules depended on when it was approved. Its review arrangements depended on who funded it and when the application was due. The institution’s systems had to support both old and new rules at once. And its budget had to absorb costs that the delay moved but did not remove.
We take the week in four clocks. Each answers a separate question, and together they cover what a research office had to manage:
- The rulebook clock: which version of the Common Rule governed a given study.
- The review clock: when a multi-site study had to use a single IRB, and under whose rule.
- The systems clock: when forms, templates and software could change to the new rules.
- The money clock: who paid for single IRB review and for running two sets of rules.
Our argument is that the institutions best placed after this week were those that tracked each study’s governing rules as data, attached to the study record. A study’s approval date, funding source and site structure decided which rules applied to it. An institution that kept those facts in its IRB system could apply the right rules automatically, through several more changes of date. One that kept them in staff memory was exposed every time a date moved. After the four clocks, we look at how this differed across institutions, at objections, and at what a research leader could check.
A note on our reading. We describe the Common Rule from the federal agencies’ Federal Register notices and the current Code of Federal Regulations, and the NIH policy from NIH’s own notices. How these interact for a given study is our reading. Institutions should rely on their own counsel, their IRB leadership and the funding agency.
The rulebook clock: which Common Rule governed a study
The first clock decided which rules applied. Under the interim final rule, the pre-2018 Requirements continued to govern all research until July 19, 2018. The rule said regulated entities could not, before that date, comply with the 2018 Requirements in place of the old ones.
No early adoption
The ban on early adoption mattered. An institution that had prepared to use the new rules on January 19 may have drafted new exemption procedures, new consent templates and new continuing review practices. The interim final rule said they had to keep using the old ones. Preparation done for January 19 had to be parked.
That created a practical problem for anything already changed. A consent template revised to the new rules, or a policy updated to end continuing review for some studies, had to be held back or reversed. An institution that had already published new policies needed to tell investigators which version applied, in writing, with a date.
Grandfathered studies
The rule set a default for existing studies. Research approved by an IRB, waived from review, or determined exempt before July 19, 2018 would stay under the pre-2018 Requirements for its duration. After July 19, an institution could choose to move such a study to the 2018 Requirements, if it determined the study would comply and an IRB documented that determination.
That made the approval date the key fact about every study. A study approved on July 18 and one approved on July 20 would run under different rules for their entire lives, unless the institution moved the first. A research office with thousands of active studies would hold two populations governed by two rulebooks, side by side.
Amendments do not reset the clock
The interim final rule tied a study’s rules to the date of its initial approval, waiver or exemption determination. In our reading, later amendments, such as new sites, new procedures or revised consent forms, did not change that date. A study approved under the old rules would be amended under the old rules, unless the institution chose to transition it after July 19. Reviewers handling amendments needed to know this, because an amendment submitted in 2019 might still be governed by the pre-2018 Requirements.
Telling investigators which rules apply
Investigators and study staff write protocols and consent forms, and they need to know which rules to write to. After January 22, an investigator preparing a new submission needed to know that the old rules still applied, and might keep applying for the life of the study. A research office that had already trained staff on the new rules had to retrain them on the old ones, or at least tell them which training to set aside.
The communication itself needed a record. A dated notice to investigators, stating which rules governed new submissions and existing studies, gave the institution something to point to if a study was later questioned. Without it, the institution would struggle to show that its staff had been told.
Exemption decisions made by staff
Where an institution lets trained staff, rather than the convened IRB, make exemption determinations, those staff apply the exemption categories in force. The delay meant applying the old categories, not the expanded ones the new rule offered. Each determination needed to record the rule version used, because the categories differed and the study would keep that determination.
A moving target
The interim final rule also announced that the agencies were developing a proposal to delay further, to January 21, 2019. So the July date was itself uncertain. An institution planning for July 19 knew it might move. Planning for a date that might not hold has a cost. Not planning for it has a different cost if it does hold.
What changed in the new rulebook
The revised rule made several changes an institution would have to apply study by study. It revised the definition of research to exclude four kinds of activity: certain scholarly and journalistic work, certain public health surveillance, certain criminal justice activities and certain national security activities. It added and revised exemption categories, some requiring limited IRB review and some relying on broad consent. It removed continuing review for some minimal risk research. It required consent forms to begin with a concise presentation of key information.
Each change applied only to studies under the 2018 Requirements. An IRB reviewer working on a given day needed to know which rules governed the study in front of them before applying any of these provisions.
Federal agencies’ own studies
The Common Rule binds federal departments and the institutions they fund. Each signatory agency codified it in its own part of the Code of Federal Regulations. The interim final rule amended each agency’s codification. An institution funded by several agencies had to confirm that each one’s version carried the same dates, and that none had diverged.
The review clock: single IRB under two different rules
The second clock governed who reviewed multi-site research. Two federal requirements pointed the same way, on different dates and with different scopes.
NIH’s policy, from January 25, 2018
NIH’s single IRB policy applied to competing grant applications with due dates on or after January 25, 2018. It covered domestic sites of NIH-funded multi-site studies in which each site conducted the same protocol involving non-exempt human subjects research. It applied to grants, cooperative agreements, contracts and NIH’s intramural program. It did not apply to career development, research training or fellowship awards, or to foreign sites.
The policy put duties on several parties. Applicants had to submit a plan describing the single IRB of record. Awardees had to ensure authorization agreements were in place, keep copies, and set up communication between the single IRB and the participating sites. Participating sites remained responsible for obtaining consent, overseeing the protocol locally and reporting problems to the single IRB. They also had to give the single IRB the local context and state requirements it needed.
The Common Rule’s mandate, from January 20, 2020
The revised Common Rule had its own single IRB requirement, in section 46.114(b) of the HHS version. It required any U.S. institution engaged in cooperative research to rely on a single IRB for the U.S. portion. The supporting agency could identify the reviewing IRB, or the lead institution could propose it subject to the agency’s acceptance. Its compliance date was January 20, 2020, and the interim final rule did not change it.
The two rules differed. NIH’s applied to NIH funding, by application due date. The Common Rule’s applied to all research covered by the Common Rule, by the 2020 date. Their exceptions differed too. NIH allowed exceptions where law required local review, or for a compelling justification. The Common Rule excepted research where law required more than one IRB, or where the supporting agency determined a single IRB was not appropriate.
What a single IRB changes
Under a single IRB, a participating site stops reviewing the protocol itself. It relies on another institution’s IRB. That shifts work rather than removing it. The reviewing IRB needs information about each site. The relying site needs a way to receive approvals, provide local information and report problems. The agreement between them has to say who does what.
NIH’s 2016 notice summarized public comments on what a single IRB needed. One comment listed capabilities such as knowledge of state law and local standards, systems for collecting information from participating sites, ways to assess site investigators, and ways to obtain community information for sensitive research. Those are systems and staffing requirements for the reviewing IRB.
Institutions as reviewers and as relying sites
Large research institutions would be both. They would serve as the single IRB for studies they led and rely on others for studies they joined. Each role needed its own procedures. As reviewer, the institution needed capacity to oversee sites it did not run. As relying site, it needed to stop duplicating review while keeping its own obligations, such as conflict of interest checks and ancillary reviews, in place.
The NIH policy did not forbid a site from duplicating the single IRB’s review. It said NIH funds could not pay for the duplicate review. That put a price on local re-review without banning it.
Who chooses the single IRB
Under NIH’s policy, the applicant described in its plan the single IRB that would serve as IRB of record for all sites. The funding institute or center managed the award and would work with the awardee to resolve questions about the IRB selected. Under the Common Rule’s provision, the supporting agency could identify the reviewing IRB, or the lead institution could propose one for the agency to accept.
In both cases, the choice was made early, at application or study start, and it shaped every site’s work for the life of the study. A lead institution choosing its own IRB took on the reviewer’s role for every site. Choosing an independent IRB moved that role outside, at a price.
Local context without local review
A single IRB reviews for sites it does not know. NIH’s policy said participating sites must give the single IRB the local context and state or local regulatory requirements it needs. That turns local knowledge into a document. Age of majority, mandatory reporting rules, language needs of the local population and site capabilities all have to reach the reviewing IRB in writing.
For a relying site, that is a new deliverable. The site’s research office had to collect the information, keep it current and send it for each study. A site that did this badly would see the single IRB approve consent forms that did not fit its population or its state’s law.
Exceptions as a process
NIH’s policy allowed exceptions where federal, tribal or state law, regulation or policy required local review. It would consider other exceptions with a compelling justification, and NIH would decide. An institution that believed a study needed local review had to make that case to NIH, in the application, with reasons. That was a new process for research offices, with its own timing.
Reliance agreements as records
Each multi-site study needed an authorization agreement, also called a reliance agreement. NIH’s policy defined it as the document setting out authorities, roles, responsibilities and communication between the reviewing IRB and the relying site. An institution in dozens of multi-site studies would hold dozens of agreements. Keeping them findable, current and linked to the right studies was a records task with compliance consequences.
The systems clock: forms, templates and software for two rulebooks
The third clock governed the institution’s tools. The interim final rule itself named this clock. Its good cause explanation cited the need to revamp institutional procedures and electronic systems, and said regulated entities might need new policies, procedures and information technology systems.
What had to change
The revised rule touched a wide range of the artifacts an IRB office uses. Application forms needed new questions for the revised exemptions and the research definition carve-outs. Consent templates needed a key information section. Continuing review schedules needed new logic for studies that no longer required it. Exemption determination tools needed the new categories, including those requiring limited IRB review.
Each of those changes had to be made in a way that left the old version available. Studies under the pre-2018 Requirements still needed the old forms and logic. An IRB system that could only hold one version of a form had to be extended, or the office had to run two sets of documents by hand.
Guidance that did not yet exist
The interim final rule said guidance from the Common Rule agencies would help implementation, and its analysis said significant guidance was needed in several areas. It named the new exemption categories, limited IRB review, broad consent, the exemption for certain HIPAA-covered activities, and which federal non-research information collections qualified for exemption. Building systems before that guidance arrived risked building them twice.
That was a reason for the delay. It was also a reason for institutions to separate the parts of their systems that depended on guidance from those that did not. A key information section in consent templates did not depend on new federal guidance in the same way that a broad consent process did.
Version control for rules
The deeper systems requirement was to attach the governing rules to each study as a data field. A study record that carried its approval date, its applicable rule version and any later decision to transition could drive the right forms, the right review schedule and the right consent template. That turned the delay from a manual sorting problem into a configuration change.
The current regulation shows why this mattered. Its transition provisions now refer to January 21, 2019, and to a documented decision to transition between July 19, 2018, and January 20, 2019. A system that stored the rule version per study could absorb those changes. One that did not would need people to sort studies by hand again.
Training reviewers on two rulebooks
IRB members and staff reviewers had to apply whichever rules governed the study in front of them. A reviewer who had trained on the new exemption categories, or on the key information requirement, had to set that training aside for pre-2018 studies and apply it for new ones after the compliance date. Review checklists keyed to the rule version, and shown automatically by the system, reduce the chance of applying the wrong one.
The audit trail
Federal oversight of human subjects protection looks at IRB records. The regulation requires IRBs to keep documentation of their activities, including research proposals reviewed, minutes and continuing review records. Under a dual rulebook, the records had to show which rules each decision applied. An IRB that could not show the rule version behind a determination would struggle to defend it.
When the dates move again
Each change of date changed which studies fell under which rules. A system that computed a study’s rule version from its approval date, using a configurable cut-off, could absorb a new date by changing one setting. A system that stored a fixed label typed in by staff would need each record reviewed. The interim final rule announced that another change was being proposed, so this was a foreseeable need.
Clinical trial consent posting
The revised rule added a requirement to post one IRB-approved consent form for each clinical trial conducted or supported by a federal department or agency, on a publicly available federal website. That needed a process: which form, by whom, when, and where. Like the other provisions, it applied to studies under the 2018 Requirements. A grants and contracts office needed to know which of its trials it applied to.
The money clock: who paid for the change
The fourth clock concerned cost. The delay moved costs in time. The single IRB policy moved costs between institutions. Both had budget consequences for research offices.
The delay’s own accounting
The interim final rule’s analysis treated the delay as postponing both costs and benefits. It estimated cost savings from not having to implement the new rules for six months. It also estimated foregone benefits, such as the burden relief institutions would have enjoyed from new exemptions, the end of IRB review of grant applications and the end of some continuing review. Its figures showed the foregone benefits as larger than the savings.
The agencies’ own analysis assumed that half of regulated entities would not have used the new or expanded exemptions during the delay, because guidance was missing. That assumption was the agencies’ estimate, not a measurement. It shows that the burden relief depended on guidance as much as on the effective date.
Paying for single IRB review
NIH’s policy allowed the additional costs of single IRB review to be charged to grants or contracts as direct costs, with appropriate justification, if reasonable and consistent with cost principles. It pointed to a separate notice, NOT-OD-16-109, on direct and indirect cost treatment. That made the single IRB a line in the budget. A reviewing institution could recover its costs from the grant. A relying institution saved its own review costs, if it did not duplicate.
The budgeting had to happen at the application stage. An application due on or after January 25, 2018, needed a single IRB plan and, where costs were claimed, a budget for them. Investigators preparing applications that month needed the research office’s help to price a service their institution had not priced before.
Applications on either side of the date
An application due before January 25, 2018, did not fall under NIH’s single IRB policy. One due on or after that date did. Two otherwise similar multi-site studies, submitted weeks apart, would be awarded under different expectations. Research offices had to check due dates, not submission dates or award dates, to know which applied, since the NIH notice keyed the policy to application due dates.
Running two systems
The dual rulebook had a cost the agencies’ analysis did not isolate: running two sets of forms, templates and procedures in parallel. That cost fell on IRB offices, research administration and IT. It would continue for as long as pre-2018 studies remained active, and long studies stay active for a long time. Institutions that chose to transition older studies to the new rules could shorten that period, at the cost of reviewing each one.
Contracting for single IRB services
An institution relying on an independent IRB, or on another institution’s IRB, entered a service relationship. Fees, turnaround times, responsibilities for reporting and the handling of disputes all needed terms. A reliance agreement covers regulatory roles. A commercial engagement may need a contract as well. Research offices that had never bought IRB review as a service had to learn to specify and price it.
Who bore the cost of waiting
The delay gave institutions time, which they had asked for. It also left investigators under rules the agencies themselves described as more burdensome in some respects. Researchers with minimal risk studies, who would have been relieved of continuing review, kept it for longer. The cost of waiting fell on them, in time spent on reviews the new rules would not have required.
How this differed across institutions
The four clocks ran in every institution subject to the Common Rule. Their effect depended on size, funding mix and role in multi-site research.
Academic medical centers
Large academic medical centers carry large study portfolios and lead multi-site work. They faced every clock at full scale. They also had the capacity to serve as single IRBs for others, which brought revenue and responsibility. For them, the reliance agreement inventory and the per-study rule version were central.
Community hospitals and health systems
Health systems that participate in trials led elsewhere would mainly be relying sites. Their work shifted from reviewing protocols to supporting a remote IRB: local context, conflict of interest, ancillary reviews and site oversight. A health system that had relied on external IRBs before could extend that process. One that had not needed to build it.
Independent IRBs
Independent IRBs, which review research for other organizations, were natural single IRBs. The NIH policy and the coming Common Rule mandate could add demand for their services. They also needed to handle two rulebooks for their clients, by study. An independent IRB serving dozens of institutions would see each client’s choices about transitioning older studies. Its own systems had to record, for each study, the rules in force and the client’s decision. Its contracts had to say which party was responsible for telling the other when that decision changed.
Life sciences sponsors
Industry-sponsored research regulated by the Food and Drug Administration follows FDA’s own human subject and IRB regulations. The Common Rule delay did not change those. But a study that was both federally funded and FDA-regulated had to satisfy both sets of rules. Sponsors working with academic sites would see the sites’ processes change around them, on the sites’ clocks. A sponsor running a multi-site trial would need to know which sites applied which rules, and which IRB each relied on, to plan its own submissions and timelines.
Social and behavioral researchers
Researchers in the social and behavioral sciences, at universities and in health systems, conduct minimal risk research such as surveys and interviews. The new exemption categories and the end of continuing review for some minimal risk studies were aimed in part at their work. The delay kept them on the old process for longer. For a university research office, that was a cost borne largely by departments outside medicine.
Smaller research sites
Smaller sites, such as community practices and clinics that join trials, have little research administration. For them, a single IRB could reduce work, since they no longer needed their own review. It also meant depending on a lead institution or an independent IRB for approvals, and meeting that IRB’s reporting requirements. Their main need was a clear point of contact and simple templates for local information.
Federal intramural research
NIH’s single IRB policy applied to its own intramural program, for multi-site studies submitted for initial review after its effective date. Federal agencies that conduct research themselves were subject to the Common Rule as signatories. They faced the same clocks as their grantees, inside their own organizations.
Public health agencies and state universities
State and local public health agencies and state universities conduct federally funded research and run surveillance programs. The revised definition of research carved out certain public health surveillance activities by a public health authority. That carve-out, once in force, could move some activities out of IRB review entirely. Until then, the old definition applied. A health department planning surveillance work in early 2018 had to route it through the old process, and keep a record of the decision, knowing the answer might differ under the new definition later that year or the next.
Objections, in order
We take five objections, numbered in the order we address them.
1. A six-month delay changed nothing important
The first objection is that the delay simply postponed a known change. It did more. It barred early adoption, so prepared institutions had to stand down. It fixed the approval date as the key to each study’s rules. And it signaled a further delay, so the new date was uncertain from the start. Each of those affected how institutions spent the following months. An office that had spent the autumn preparing for January 19 lost the use of that work for at least six months, and then had to prepare again for a date that was itself provisional.
2. NIH’s policy and the Common Rule’s mandate were the same thing
The second objection is that the two single IRB requirements were one idea. They shared a purpose. They differed in scope, start date, exceptions and enforcement. NIH enforced its policy through award terms. The Common Rule’s mandate applied as regulation, from 2020. A study could fall under one, both or neither, and the institution had to know which. A study funded by NIH and another agency, with sites in several states, might face NIH’s policy at application and the Common Rule’s mandate later, with different exception rules at each stage.
3. IRB software vendors would handle the dual rulebook
The third objection is that IRB systems vendors would update their products. Vendors could add new forms and logic. Whether a given institution’s configuration attached the right rule version to each study depended on the institution. A vendor cannot know which studies an institution chooses to transition. Vendors also release on their own schedules. An institution waiting for a vendor release before acting would be on the vendor’s clock as well as the federal ones.
4. Investigators, not offices, carry research compliance
The fourth objection is that investigators are responsible for their studies. They are, for conduct. But which rules apply, which forms to use, and which IRB to rely on are decided by institutional systems and offices. An investigator cannot reasonably track four federal clocks while running a study. The institution’s systems can, and that is where the duty to get it right in practice falls. Pushing the tracking onto investigators would spread the same work across hundreds of people, each with less context than the research office.
5. The burden relief was small, so the delay cost little
The fifth objection is that the new rules’ relief was modest. The agencies’ own analysis put the foregone benefits of the six-month delay at 49.5 million dollars a year at a 3 percent discount rate, against cost savings of 7.4 million. Whatever the precise figures, the relief was concentrated on minimal risk research, where investigators spend IRB time on studies that pose little risk.
What a research leader could check
These checks follow from the four clocks. They are our suggestions, not requirements.
- Does every active study record carry its governing Common Rule version, its approval date and any documented decision to transition?
- Does the IRB system serve the correct forms, templates and review schedules from that field, rather than relying on staff to choose?
- Is there a current inventory of reliance agreements, linked to studies, showing which institution reviews and which relies?
- Are single IRB costs priced and included in applications where the funder allows them as direct costs?
- Does the institution track which studies fall under NIH’s single IRB policy, the Common Rule’s cooperative research provision, both or neither?
Each check can also be measured. Useful figures include the share of active studies with a rule version recorded and the number of multi-site studies with a current reliance agreement on file. The share of applications with a priced single IRB plan is another. Those figures tell a research leader whether the institution’s records will absorb the next change of date.
The checks need owners. The rule version field belongs to the IRB office, with the IT team that runs the system. Reliance agreements belong to the human research protection program. Single IRB budgeting belongs to research administration and the office of sponsored programs. The tracking of funder policies belongs to whoever reads federal notices for the institution, which should be a named role.
With hindsight
The further delay came. On June 19, 2018, the Common Rule agencies published a final rule delaying the general compliance date of the 2018 Requirements to January 21, 2019. Its summary says the January 2018 interim final rule had been issued on January 17, 2018, before publication on January 22. It also notes that the Department of Housing and Urban Development published its own interim final rule adopting the same changes on January 26, 2018.
The June rule did something the January rule had forbidden. It allowed institutions to use three burden-reducing provisions of the 2018 Requirements during the delay, from July 19, 2018. The current regulation lists them: the revised definition of research, the removal of IRB review of grant applications, and certain exceptions to continuing review. An institution had to document and date a determination to transition a study to use them. The compliance date for the cooperative research provision remained January 20, 2020.
The current text of 45 CFR 46.101(l) now carries the full set of clocks. Research initially approved, waived or determined exempt before January 21, 2019 stays under the pre-2018 Requirements unless the institution transitions it. Research approved on or after that date follows the 2018 Requirements. A study transitioned between July 19, 2018, and January 20, 2019 followed a defined mix. That is consistent with our first clock: the approval date and any documented transition decide a study’s rules for its life.
The single IRB requirement arrived on schedule. NIH’s notice of January 22, 2020, told the research community that NIH would adhere to the revised Common Rule’s single IRB requirement for multi-site studies. It said that from January 20, 2020, NIH-funded multi-site studies supported by career development and fellowship awards, which NIH’s own policy had excluded, would be subject to 45 CFR 46.114 if they obtained initial IRB approval on or after that date. It also described the Office for Human Research Protections’ November 2019 determination excepting two categories of research from the single IRB requirement. Both categories turned on decisions made before January 20, 2020.
That notice fits our second clock. NIH’s policy and the Common Rule’s provision were different instruments with different scopes, and NIH had to explain how they fit together.
FDA moved toward the same model. On September 28, 2022, it proposed a rule that would require U.S. institutions in FDA-regulated cooperative research to rely on a single IRB, with some exceptions. The proposal said it would harmonize FDA’s requirements, as far as practicable, with the revised Common Rule, and that it was issued under a provision of the 21st Century Cures Act. It also proposed an IRB recordkeeping requirement for research at an institution whose IRB oversight is conducted by an IRB the institution does not operate. Comments were due November 28, 2022.
For a research leader today, the practical point is that federal research rules change by date, and studies outlive the dates. Every future revision, whether to the Common Rule, FDA’s rules or a funder’s policy, will create another population of studies under older terms. Storing the governing rules per study is the one practice that makes each new change a configuration task rather than a sorting exercise.
What held up is that research compliance became a matter of tracking each study’s governing rules over time. What the week did not show is how long the dual rulebook would last. The current regulation still preserves the pre-2018 Requirements for studies approved before January 21, 2019, which means long-running studies can still be governed by rules first written in 1991.
Spotted an error? Tell us and we will check it and correct the article where needed.