NIH draft requires safety plans for all clinical research

By Rebecca Stone

NIH would require every NIH-supported clinical research study—including observational work—to have an approved safety-monitoring plan before human-subject activity begins, under a draft policy released Nov. 2, 2026.

The proposal would also make independent-monitoring requirements depend on study risk. It would remove categorical requirements for a data and safety monitoring board (DSMB), an independent study-monitoring body, for Phase 3 or multisite trials.

NIH said the update seeks to “cut administrative burden that doesn’t match a study’s risk while ensuring reliable data.” In the announcement, NIH presents that as an aim rather than a measured result.

What does the draft cover?

The draft would revise the 1998 NIH Policy for Data and Safety Monitoring, which, together with 2000 guidance on monitoring Phase 1 and Phase 2 trials, has set the baseline for NIH-funded trial oversight.

The 1998 policy required monitoring plans only for clinical trials. The proposal would extend that requirement to all NIH-supported clinical research: epidemiological and behavioral studies, outcomes research, health services research, and studies of disease mechanisms.

Coverage would include competing grant applications, competitive revisions, contract proposals, other funding agreements, and intramural clinical research projects approved by an institutional review board (IRB). The funding institute, center, or office would approve the plan.

This broadens the pre-award workflow. In grant and contract planning, investigators would need to define the monitoring approach before any human-subject work starts, rather than build that step around trial protocols alone.

Why do current requirements vary?

The 1998 policy allowed each NIH institute to implement monitoring as it saw fit. NIH-wide grant instructions require a formal data and safety monitoring plan for clinical trials but make it optional for other human-subject research, including observational studies.

Institute practice already differs. NIDDK requires monitoring plans for observational studies. NIAMS says large observational studies may need their own monitoring board.

An analysis of ClinicalTrials.gov records shows wider variation in reported arrangements. It covered registered Phase 3 trials, including combined Phase 2/3 trials, with actual or estimated start dates from 2017 onward.

Among trials answering the data monitoring committee (DMC) question, the reported shares were:

• NEI: 100%, 20 of 20 trials.
• NIAID: 98%, 43 of 44 trials.
• NICHD: 96%, 49 of 51 trials.
• NIA: 91%, 29 of 32 trials.
• NCI: 50%, 135 of 272 trials.

The analysis used registry data retrieved Nov. 5, 2026. It covered interventional studies listing an NIH organization as lead sponsor or collaborator, included future start dates, and included every trial status, including withdrawn and terminated studies.

Registrants supplied the DMC answers, and the percentages excluded unanswered fields. A study could count under more than one institute. The figures establish reported committee appointments, not compliance with NIH requirements, so they cannot measure how many plans or boards the agency actually required.

What changes for non-trial studies?

The extension to non-trial research is the draft’s largest operational change. NIH says almost two-thirds of its clinical research portfolio consists of studies that are not clinical trials, bringing a majority of the portfolio into the draft’s formal plan-review framework for the first time.

NIH says data from these studies often feed into later trials. The agency wants the same data-quality and safety standards to apply across both stages, although the proposal gives institutes discretion to scale oversight according to risk.

For R&D managers, the proposal would change workload assumptions before award activation. Portfolio teams would need monitoring plans and institute, center, or office approval across studies that previously fell outside the trial-only requirement.

What does the approval timetable mean?

Each in-scope study would need an approved plan before human-subject activities begin. The draft sets a 30-day response period for the relevant institute, center, or office after receiving a requested plan; after that period, the plan would be deemed approved.

Researchers would have to report amendments. Changes covered by NIH’s prior-approval rules would need approval before implementation, adding another checkpoint to protocol and operational changes.

NIH would check compliance at regular reporting intervals. Noncompliance could lead to added award conditions or termination, so investigators would need plan documentation and amendment histories available throughout the award period.

The draft would also apply to intramural work, bringing grant-supported investigators and NIH’s internal research programs under the same plan-review framework.

When would the policy take effect?

NIH is accepting comments through Jan. 6, 2027. A final policy would apply only to applications submitted at least six months after its publication, while existing requirements would remain in force during the transition.

The agency wants to know whether the draft improves risk-based oversight of data quality and participant safety. It also seeks proposals for compliance resources, including templates, model monitoring plans, a risk-assessment framework, and guidance on required plan contents.

NIH said it wants to “harmonize monitoring requirements across the agency.” The proposal treats harmonization as a remedy, although the registry figures show substantial variation in reported committee arrangements, particularly between NCI and the other cited institutes.

Separately, NIH sought comment in June on a proposal to cap simultaneous research-project grants per investigator. The source also notes publication-cost concerns surrounding NIH’s public-access policy. Those policies do not alter the monitoring draft, but grant caps and publication costs still enter awardee portfolio and budget planning.

The next checkpoint is the Jan. 6, 2027 comment deadline; whether the final rules reduce duplicated oversight or extend compliance work to most of the portfolio will depend on the risk framework, templates, and six-month implementation window NIH is still developing.