NIH Scientific Impact RFI (NOT-OD-26-087): What It Means for Your Next Grant

What NOT-OD-26-087 Is Actually Asking

The RFI frames a question the biomedical community has argued about for years: is impact score enough? NIH has traditionally relied on peer review scores, citation counts, and downstream grant productivity to judge whether funded science "worked." That approach has real weaknesses. A lab that produces replicable, well-documented, widely shared results contributes differently than one that publishes striking findings no one can reproduce — and the current scoring system does not reliably distinguish between them.

NOT-OD-26-087 organizes the question into three distinct buckets. First, reproducibility and replication: NIH wants to know how it should identify fields where replication studies are most needed, and how it might reward investigators for pursuing them rather than only chasing novel results. Second, data, software, and model sharing: NIH is asking explicitly how it should measure and reward sharing behavior, including through a Data Sharing Index (S-index) that NIH is currently piloting as a challenge. Third, training and mentorship: NIH is exploring whether formal metrics for mentorship quality and trainee outcomes could become part of how career-stage investigators are evaluated for future awards.

This is not NIH's first pass at any of these themes. NIH launched a central resource for replication and reproducibility in early 2026, the 2025 Simplified Peer Review Framework already folded rigor requirements into Factor 2 review, and the 2026 DMS plan revision made sharing plans shorter but more intent-focused. What is new about NOT-OD-26-087 is the explicit ask: tell us how to quantify these things so we can reward them. That framing suggests NIH is moving toward embedding these metrics more directly into funding decisions rather than treating them only as review criteria. The timeline is uncertain, but the direction is not.

Reproducibility in Current Review — and Where It Could Go

The four NIH rigor and reproducibility requirements — scientific premise, rigor of prior research, biological variables, and authentication of key resources — are already scored under Factor 2 of the Simplified Review Framework. Reviewers are trained to flag applications that do not address all four clearly in the Approach section. In practice, many reviewers still treat these as checkbox items unless there is an obvious gap, so a well-written Approach that addresses all four without drawing attention to itself usually holds up in review.

What the RFI signals is that this may shift over time. If NIH develops formal metrics for reproducibility practices — pre-registration, statistical power documentation, protocol sharing, independent validation of key reagents — those metrics could eventually feed into reviewer training or scoring rubrics. We don't know the timeline. RFIs don't commit NIH to any particular action, and implementing new review criteria takes years. But the direction is clear enough that adjusting your applications now costs nothing and could differentiate you if scoring weights change.

For your current Approach section: if you rely on cell lines, antibodies, or animal models that have documented reproducibility problems in the literature, name them and explain how you will address that concern. If your preliminary data were generated with specific statistical approaches, state your power assumptions explicitly — not buried in methods, but visible in the argument. These choices help reviewers do their job and position you well for a world in which reproducibility is scored more directly.

The S-Index and What Data Sharing Metrics Signal for Your DMS Plan

The Data Sharing Index (S-index) is a challenge NIH launched to explore how to quantify data sharing impact. The concept: a researcher who shares high-quality datasets in accessible repositories with good documentation should get credit for that in some trackable way, comparable to how citation metrics credit publication impact. The S-index is experimental — it does not currently appear in peer review or any application form — but its presence in NOT-OD-26-087 signals that NIH wants sharing behavior to eventually matter in a measurable, documented way.

For your Data Management and Sharing plan, the 2026 format already simplified required elements to three core yes/no questions. What the S-index signal adds is this: the quality and accessibility of how you share data is starting to be visible in ways it was not before. A DMS plan that names a well-documented public repository, explains why a researcher in another field could actually use your data, and addresses reuse rather than just deposit reads differently than a compliance-minimum plan. Reviewers who care about data sharing — and more of them do each cycle — read the DMS plan more carefully than the form's length implies.

You don't need to mention the S-index in your application. But write your DMS plan as if someone will eventually check whether you did what you said. Choose repositories that are genuinely accessible. Describe any data dictionaries or codebooks the data requires. Be honest about restrictions and explain why they exist. That kind of plan will survive whatever metric NIH eventually adopts.

Mentorship Metrics: Implications for K Awards and T32 Applications

The mentorship and training bucket of NOT-OD-26-087 is the most speculative, but it is also the one that could most directly reshape K award and T32 review if it leads to policy. Career development award applications already require mentor qualifications and a list of prior trainees. T32 applications require detailed tables of trainee outcomes — publications, job placements, funding records. The RFI asks whether these should be formalized into measurable scores rather than evaluated as narrative.

If NIH introduces a formal mentorship quality metric, it will almost certainly draw on trainee outcome data already tracked in xTRACT and similar NIH systems. That means the mentorship record your current trainees are building — publications, fellowship applications, career outcomes — is already the input for any future metric. You cannot retroactively improve your historical record, but you can document it more carefully right now. In K award applications, adding a brief table of prior trainees with outcomes before it is required signals seriousness and gives reviewers something specific to evaluate rather than a list of assertions.

For T32 applicants and directors, this is a useful moment to audit your xTRACT data. T32 renewal applications that lack clean outcome records for prior trainees consistently draw reviewer concern, and that concern is likely to grow as NIH signals it wants to reward training impact more explicitly. If your program has strong trainee outcomes scattered across individual CVs rather than assembled anywhere central, fix that now rather than the month before your renewal deadline.

Should You Respond to the RFI?

The deadline for NOT-OD-26-087 comments is August 19, 2026. Responding is not required and has no effect on any application. Whether it is worth your time depends on what you actually have to say.

Researchers who have specific, evidence-based views on any of the three buckets are exactly the audience NIH is trying to reach: PIs who have published replication or reproducibility studies and can speak to incentive gaps; investigators who have shared datasets with documented downstream reuse; T32 directors with structured trainee outcome tracking. A two-paragraph comment grounded in a concrete experience is more useful to NIH than a longer abstract opinion. If you don't have a strong evidence-based view on any of these areas, a busy submission cycle is probably not the right time to draft one. Check the NOT-OD-26-087 page on grants.nih.gov for the submission link and the exact question prompts before deciding.

What to Do in Your Next Application Right Now

None of the changes signaled by NOT-OD-26-087 are in effect yet. Peer review rubrics have not changed. The S-index is not in applications. Mentorship metrics are not a scored criterion. But adjusting your applications now costs almost nothing and could matter as policy evolves.

These are not radical changes. They are the kind of documentation that makes a strong application stronger and gives reviewers the specific evidence they need to advocate for you at the panel. The RFI is a signal about direction. Treating that signal as an early warning — rather than waiting for policy that may be two or three years away — is a reasonable grant strategy at any career stage.

Frequently Asked Questions