NIH Just Replaced the Data Management & Sharing Plan With Six Yes/No Questions. The May 25 Deadline Has Already Passed — Here Is What Changed and Why It Matters.

When the NIH Data Management and Sharing Policy went into effect in January 2023, it was sold as a modest administrative addition — a one-to-two-page plan describing how researchers would share the scientific data underlying their grant-funded work. Three years and roughly 200,000 plans later, what emerged was something nobody on either side of the review process actually wanted: long, vague, copy-pasted narratives that obscured rather than clarified data-sharing commitments, evaluated by reviewers who often could not quickly tell whether the plan was specific enough to enforce.

On May 25, 2026, the NIH retired that format. Guide Notice NOT-OD-26-046 — issued February 25, 2026 with a 90-day implementation window — replaces the narrative DMSP with a structured checklist. All applications with due dates on or after May 25 must use the new format. The 2023 format is no longer accepted. For research administrators and principal investigators who built institutional templates around the old structure, the practical implication is immediate: every active grant proposal in the pipeline needs a DMSP update before it goes out the door.

The change is unusual for NIH. The agency typically extends transition periods, allows parallel formats, and gives the research community quarters or years to adapt. The 90-day transition here reflects a different calculation: the 2023 format was producing plans that the agency could not effectively monitor for compliance, and dragging the transition out longer would have meant another year of plans that nobody could enforce. The new format is designed primarily to make compliance auditable. It happens, as a side benefit, to be much faster to write.

What the New Format Actually Looks Like

The 2023 narrative format required researchers to address six elements in continuous prose: data types and amount, related tools and code, standards, data preservation and access, oversight, and special considerations like human-subjects protections. The result was typically a two-page document that read like a research methods chapter, with sections that often duplicated information already provided elsewhere in the application.

The 2026 format compresses this to three components: a structured set of yes/no questions addressing the core sharing commitments, a single data types and repositories table capped at 100 words, and brief text explanations required only when an answer is "no" or "limited." The three core yes/no questions ask whether scientific data underlying peer-reviewed publications and other findings resulting from the grant will be shared, whether the data will be shared by required timelines, and whether the data will remain available for as long as repository or journal policies require. A 300-word explanation is required only if the answer to any of these is qualified by ethical, legal, or technical limitations.

The shift from narrative to checklist sounds bureaucratic, but the substantive consequence is meaningful. A 2023-format plan could be technically compliant while remaining functionally vague: a researcher could state intent to share, name a generic repository, and never commit to specific timelines or access mechanisms. A 2026-format plan forces a binary answer to whether data will actually be shared, by when, and for how long. The auditor — whether an NIH program officer, an institutional research integrity office, or a journal editor — can determine compliance in seconds rather than minutes, and the researcher cannot hide behind narrative ambiguity.

Why the 2023 Format Failed

The 2023 DMSP was the product of nearly a decade of NIH internal debate about how to operationalize data sharing without creating a compliance burden that would suppress applications. The compromise was a narrative format that gave researchers flexibility to describe their plans in field-appropriate terms — a genomics researcher and a clinical psychologist could each describe a sharing plan that fit their data type and discipline.

The flexibility worked against the policy goal in practice. The narrative format invited padding. Researchers copied repository descriptions from institutional templates, included generic statements about FAIR data principles, and used phrases like "data will be made available upon reasonable request" that had no enforceable meaning. Reviewers, asked to evaluate scientific merit on dozens of pages of proposal text, did not have the bandwidth to scrutinize a two-page narrative for compliance with policy requirements they could not enforce anyway. The result was a system where plans were written, plans were approved, and the actual data sharing behavior of funded researchers diverged from the plans in ways that the agency could not detect.

The accelerated NIH multiyear funding shift documented in our coverage of the $402 million FY2026 reallocation is one factor pushing the agency toward administrative simplification across the board. With fewer competing grants and more scrutiny on each award, the NIH needs compliance mechanisms that actually work. A simplified DMSP that the agency can audit at scale is a small but real piece of that broader administrative tightening.

The other factor is the broader 2 CFR 200 Uniform Guidance revision and the Executive Order 14332 federal grant oversight regime implemented in 2025. The new oversight framework expects agencies to demonstrate that grant compliance requirements are not just nominal — that the documents grantees submit can actually be checked against grantee behavior. A narrative DMSP that nobody reads is exactly the kind of theater that the new oversight regime is designed to eliminate.

What Researchers Need to Do Right Now

For active applications with due dates on or after May 25, the DMSP must be rewritten in the new format. The 2023 plan cannot be submitted as-is, and a 2023 plan converted to the new format by simply pasting content into the new fields will not work — the new format demands binary answers that the old narrative format was designed to avoid. The realistic time investment for a researcher with an existing 2023 plan is 30 to 60 minutes per application to translate the substantive commitments into the structured checklist format.

The 100-word data types and repositories table is the place where most translation effort will concentrate. Researchers need to identify, for each major data type the grant will produce, the specific repository where it will be deposited and the access conditions that will govern that deposition. Generic answers — "appropriate repositories will be identified during the project period" — will not survive review. The reviewer needs to see a specific named repository (dbGaP, OpenNeuro, ImmPort, GEO, ICPSR, etc.) and a specific access mechanism (open, controlled, embargoed-then-open).

The 300-word explanation when a sharing commitment is qualified is where strategic writing matters. Researchers working with human-subjects data, indigenous community data, or proprietary industry data have legitimate reasons to limit sharing, but the explanation needs to make the limitation specific rather than gestural. "Data will be limited due to privacy concerns" is too vague. "Individual-level genotype data will be deposited in dbGaP with controlled access requiring an approved Data Access Request; aggregate summary statistics will be deposited in GWAS Catalog with open access" is the level of specificity that the new format expects.

Common errors that will trigger reviewer or program officer flags include: inconsistency between the DMSP and the data availability statement in the eventual publications (the repository named in the plan must match the repository named when the paper is published); promising deposition in a repository that does not accept the relevant data type; specifying timelines that conflict with journal embargo periods or repository policies; and treating the plan as separable from the rest of the application rather than integrating it with the human-subjects protections, biosafety considerations, and resource sharing sections.

The Institutional Implications

For research administrators and sponsored programs offices, the May 25 deadline triggers a template overhaul that needs to happen quickly. Most institutions built DMSP templates and boilerplate language during 2022 and 2023 to help PIs comply with the new policy. Those templates are now obsolete. Institutions that do not update their templates within the next several weeks will see their PIs submitting non-compliant DMSPs that get returned for correction, delaying applications and creating administrative friction at exactly the moment when the federal grant pipeline is already constrained by reduced opportunity counts.

The institutional response should include: updated DMSP templates that match the 2026 format with placeholders for the structured checklist responses; updated boilerplate for common data types and repositories used by faculty at the institution; training sessions for sponsored programs staff on how to review the new format; and updated guidance documents explaining the difference between the 2023 and 2026 formats. Institutions that maintain DMP-Tool integration should verify that their templates are pulling the correct 2026 format.

For PIs with multiple grants in flight, the highest-priority updates are applications with submission deadlines in June and July 2026. Earlier-deadline applications submitted before May 25 used the 2023 format and are not affected. Later-deadline applications give time to revise institutional templates and then update individual plans. The middle window — applications going out in the next 30 to 60 days — is where the time pressure is real.

The Longer Trajectory

The DMSP simplification is not the end of the NIH compliance reform process. The agency has signaled that the next phase will focus on post-award monitoring: actually checking that the data described in the DMSP is, in fact, deposited in the repositories named, on the timelines committed, with the access mechanisms specified. The 2026 format makes this monitoring possible. The 2023 format effectively did not.

Researchers who treated the 2023 DMSP as a perfunctory document — written quickly, never revisited, ignored after award — will find that the 2026 environment is different. Program officers will be able to check, in real time, whether a grant's data deposition matches its plan. Audits will be faster and more frequent. Renewal applications will increasingly require demonstration that prior plans were actually executed. The plan you write for a 2026 application is the plan you will be held to throughout the award.

For most researchers, this is not bad news. Data sharing has been a stated NIH priority for two decades, and the researchers who have built rigorous sharing practices into their workflows will benefit from a compliance regime that recognizes the difference between rigorous and performative plans. The researchers who relied on narrative ambiguity to avoid committing to specific sharing actions will need to develop the operational capacity to actually share data — which, after three years of stated policy and a now-enforceable format, is a reasonable expectation.

The May 25 deadline has passed. The applications going out next week need the new format. The institutional infrastructure that supports those applications needs to catch up to the policy change that is already in effect.