Applicants writing a Data Management and Sharing (DMS) Plan for an NIH grant, cooperative agreement, or R&D contract are not free to structure the document however they like. NIH publishes an actual template — and, as of the 2026 revision, requires its use for most submissions. This guide walks through the current template element by element: what NIH expects in each field, how it changed in 2026, and where administrators most commonly get it wrong. For the underlying policy — who has to submit a plan, when data must be shared, how the plan is reviewed — see CASRAI’s companion guide, NIH Data Management and Sharing (DMS) Policy: Requirements, Timeline, and Compliance.
Two templates are now in circulation — use the right one
NIH’s original DMS Plan format, in place since the policy took effect on January 25, 2023 (NOT-OD-21-013), was a roughly two-page narrative addressing six elements. In April 2026, NIH released a substantially shorter 2026 Pilot DMS Plan Format (announced via the NIH Extramural Nexus News post “2026 Pilot Data Management and Sharing Plan Format Available,” formalized in Guide Notice NOT-OD-26-046, “Updated Elements of an NIH Data Management and Sharing Plan”).
The transition rule: applications with due dates before May 25, 2026 could use either the original 2023 narrative format or the new pilot format. For due dates on or after May 25, 2026, the 2026 Pilot format is required for most submissions, unless a specific Notice of Funding Opportunity (NOFO) says otherwise. Because that cutoff has already passed as of this writing, nearly every applicant preparing a plan today should be using the 2026 Pilot template, not the 2023 narrative version — check your specific NOFO to confirm it doesn’t specify otherwise. NIH itself is candid that despite the “pilot” label, the downloadable format page functions as a required template, not an optional starting point the way the 2023 version did (which allowed institutions to use their own layout as long as all six elements were addressed).
The 2023 narrative format (still relevant for legacy and non-competing awards)
The original format asked applicants to address six elements in prose, typically within about two pages:
- Data type — what scientific data will be generated, and which subset will be preserved and shared.
- Related tools, software, and/or code — whether specialized tools are needed to access or manipulate the shared data.
- Standards — what data and metadata standards will be applied.
- Data preservation, access, and associated timelines — where data will be archived and when it becomes available.
- Access, distribution, or reuse considerations — consent constraints, IP, or other factors affecting reuse.
- Oversight of data management and sharing — how execution of the plan will be monitored.
This structure still matters for two reasons: some non-competing continuations and older active awards operate under plans written in this format, and the six substantive concerns it lists (what data, what tools, what standards, when and where it’s shared, what limits apply, who’s accountable) are the same concerns the 2026 Pilot format re-asks, just in a more constrained shape.
The 2026 Pilot format: what’s different
The 2026 Pilot format restructures the same substantive ground into seven elements, most of which are answered as Yes/No/Not Applicable rather than open narrative. NIH’s stated goal is to reduce applicant burden and make plans faster to review and compare, not to reduce the underlying obligations. Per NOT-OD-26-046 and corroborating institutional summaries (UCLA Office of Contract and Grant Administration; Becker Medical Library):
- Most of the seven elements are Yes/No/N.A. checkbox-style questions covering the same ground as the 2023 elements — whether scientific data will be generated, whether it will be preserved and shared, whether standards will be applied, and so on.
- One element allows a narrative justification, capped at 300 words, specifically for explaining any limits on sharing (legal, ethical, participant-consent, or technical constraints that restrict how openly data can be released).
- One element requires a short table, capped at 100 words, identifying the anticipated data types and the repository (or repositories) where each will be deposited.
Because the exact wording of each of the seven checkbox items can be revised by NIH between guide notices, don’t transcribe a specific question list from a secondary source into your institution’s boilerplate — download the current fillable format page directly from grants.nih.gov (Grants & Funding → Write Your Application → Forms Directory — Data Management and Sharing Plan Format Page) at the time you’re preparing each application, and confirm the version against the applicable NOFO instructions.
Filling out the template: practical guidance by section
Data type
Be specific about format and volume, not just subject matter — “imaging data” is too vague; “DICOM-format MRI scans, approximately 200 GB across the award period” is usable by a reviewer and by your own data office at closeout. State plainly which data, if any, will not be shared, and briefly why (this is the substance that now flows into the 300-word justification element in the 2026 format).
Related tools, software, and code
If a specialized pipeline, proprietary instrument format, or custom script is required to open or interpret the shared data, say so and name it. If code will also be shared (e.g., via GitHub with a Zenodo-minted DOI for archival), state that explicitly — reviewers and downstream users should not have to infer it.
Standards
Name the actual metadata and data standards you’ll apply — a domain-specific schema where one exists (e.g., MIAME for microarray data, BIDS for neuroimaging), or a general-purpose schema like DataCite’s metadata schema when no domain standard fits. “Standard formats will be used” without naming one is the single most common reason reviewers or program staff send a plan back for revision.
Preservation, access, and timelines
Name the specific repository, not just a category (“an appropriate repository” is not sufficient). NIH’s stated priority order is: (1) a repository the specific Institute, Center, or NOFO requires; (2) an established discipline-specific/community repository; (3) a generalist repository only if neither of the first two exists for that data type. State the timeline in concrete terms — NIH’s baseline expectation is no later than the associated publication or the end of the award’s performance period, whichever comes first. See CASRAI’s guide on choosing an open data repository and the Data Repository entry for how to evaluate options against that priority order.
Access, distribution, and reuse considerations
Flag anything that limits open sharing — informed consent language that restricts secondary use, IP considerations, or data sensitivity requiring a controlled-access repository rather than open deposit. This is where a genomics-generating project should also note that the separate NIH Genomic Data Sharing (GDS) Policy applies alongside the DMS Policy, typically requiring deposit to a controlled-access repository such as dbGaP.
Oversight
Name who is responsible for confirming the plan is actually executed — commonly the PI, with institutional data management or research computing support named as backup. Vague oversight language (“the PI will ensure compliance” with no supporting infrastructure named) is a frequent reason program staff request revision at just-in-time.
Budgeting for the plan you write
Reasonable, allowable costs of carrying out the plan — data curation and formatting, repository deposit or hosting fees, and similar costs directly tied to commitments in the plan — can be requested in the proposal budget, generally as a distinct line item rather than folded into indirect costs, provided the costs are well justified and consistent with the institution’s normal cost-accounting practices. A DMS Plan that names specific deposit or curation costs without a matching budget line is a mismatch reviewers and program staff both notice.
Common mistakes when using the template
- Reusing a 2023-format narrative for a due date on or after May 25, 2026. Confirm which format the current NOFO expects before drafting — don’t assume last year’s institutional boilerplate is still the right shape.
- Naming a repository category instead of a specific repository. “A public repository” is not an answer NIH program staff can assess for feasibility.
- Treating the plan as boilerplate copied across every application from the same lab. Data type, volume, standards, and timeline genuinely differ by project; generic language is the most common reason a plan gets flagged for revision.
- Confusing the DMS Plan with the separate manuscript-deposit obligation. The NIH Public Access Policy governs peer-reviewed manuscripts in PubMed Central; it is an independent requirement from the DMS Plan, which governs the underlying scientific data. Meeting one does not satisfy the other.
- Leaving out budget justification for named costs. If the plan states a repository deposit fee or curation cost, the budget should reflect it.
Frequently asked questions
Where do I download the actual NIH DMS Plan template?
Directly from grants.nih.gov’s Forms Directory (Data Management and Sharing Plan Format Page), not from a secondary or institutional mirror, since NIH periodically revises the fillable form itself. Confirm which version (2023 or 2026 Pilot) your specific NOFO requires before drafting.
Do I need to use the 2026 Pilot format for every current application?
For applications with due dates on or after May 25, 2026, yes, unless the specific NOFO states otherwise. Applications due before that date could use either format during the transition window.
Is the DMS Plan template the same across all NIH Institutes and Centers?
The format page itself is NIH-wide, but individual Institutes, Centers, and specific NOFOs can layer additional expectations on top — particularly around required or preferred repositories. Always check the specific funding opportunity’s instructions in addition to the general format page.
How long should the plan be?
The 2023 narrative format was informally around two pages; the 2026 Pilot format is intentionally shorter by design, constrained by its Yes/No structure plus the 300-word and 100-word limits on its two narrative fields. Length itself was never the compliance target — completeness and specificity are.
Does a DMS Plan template exist for NSF as well?
Yes, but NSF’s structure, review treatment, and 2026 format changes differ meaningfully from NIH’s — see CASRAI’s NIH vs. NSF Data Management Plans comparison guide.
This guide reflects the DMS Plan format as revised under NOT-OD-26-046 (2026 Pilot format, required for due dates on or after May 25, 2026) alongside the original NOT-OD-21-013 six-element format. NIH periodically issues further guide notices and FAQ updates through the Office of Extramural Research; verify the current template and any NOFO-specific instructions directly on grants.nih.gov before relying on this page for an active application.
