Skip to content

PAR-25-357 NEI Secondary Data R21 Eligibility Guide

Updated 5 official sources

Quick Answer: PAR-25-357 eligibility in 2026 turns on 6 checks: vision fit, existing analyzable data, clinical trial not allowed status, variable-to-aim mapping, bias plan, and manuscript milestones.

Full answer and how we checked it

PAR-25-357 eligibility in 2026 turns on 6 checks: vision fit, existing analyzable data, clinical trial not allowed status, variable-to-aim mapping, bias plan, and manuscript milestones.

Who this is for: Vision researchers, ophthalmology and optometry investigators, biostatisticians, EHR researchers, clinical trial data users, epidemiologists, and sponsored-research offices deciding whether an existing-data vision project fits the NEI R21.

Data note: This guide uses NIH PAR-25-357 (opens in a new tab), NIH clinical trials policy (opens in a new tab), NEI clinical-trial policy notice NOT-EY-23-001 (opens in a new tab), Grants.gov listing 357979 (opens in a new tab), and Grantsights award-data signals checked on June 2, 2026.

In this guide

PAR-25-357 is not a general vision R21. According to NIH PAR-25-357 (opens in a new tab), the project should use existing data to produce insight into vision health problems, and the data should already be in a form that can support the aims.

In some NIH secondary-data proposals, the data are real but the aims are not. The strongest application does not say "we have a dataset." It says which variables answer which aims, what the data can and can't prove, and what manuscripts will be produced.

Use this guide with the live NEI Vision Secondary Data R21 program record. Then compare NIH grants complete guide, NIH R01 R21 R03 K awards comparison, best NIH grants for researchers, what is a NOFO, and the Grantsights grant database.

June 2026 Source Check

6 official-source checks should happen before a team treats PAR-25-357 as the right R21. NIH lists PAR-25-357 as an NEI R21, clinical trial not allowed, for vision-related secondary analyses using existing human data from clinical trials, epidemiologic studies, clinical research, EHR, administrative claims, surveys, registries, or patient-reported data resources. NIH also states that the NOFO must not support collection of new data.

NIH asks applicants to describe the datasets or databases, show that data are structured and reasonably cleaned, account for bias and data deficiencies, and include milestones with at least 2 manuscript submissions. Grantsights tracks the active program deadline as October 16, 2027, while the NIH NOFO lists an October 17, 2027 expiration date.

Source checkWhat to verify before writing
Trial statusNo NIH-defined clinical trial or new intervention assignment
Data statusExisting vision-related data are already structured and analyzable
Aim fitEvery aim maps to variables, time points, and outcomes that exist now
Bias planMissingness, selection, misclassification, confounding, and linkage risk are handled
ManuscriptsAt least 2 manuscript submissions are built into milestones
Better routeR01, parent R21, clinical-trial route, or data-prep work is compared

The fastest fail signal is a project that needs new recruitment, new outcome collection, or large data cleanup before analysis can begin.

Get grant deadline alerts by email

About once a week, we email a short list of open grant calls with real deadlines from different funders. It isn't limited to the programs in this guide.

No spam. Unsubscribe anytime.

6 PAR-25-357 Eligibility Stop Checks

6 stop checks decide whether this R21 fits. The project should be vision-related, secondary-data based, non-trial, and feasible within a focused exploratory scope.

Stop checkStrong R21 signalRed flag
Vision fitAim addresses vision health problemVision link is incidental
Data statusExisting data are structured and usableData collection or cleaning is the main project
Trial statusNo NIH-defined clinical trialNew intervention assignment is planned
Analysis planVariables, bias, and limitations are clearFishing expedition
Output planManuscript milestones are credibleNo publication timeline
Mechanism scaleFocused R21 scopeFull R01 or new study is needed

Don't use this route for primary data collection in disguise. The whole case depends on data that already exist and can support the proposed analysis.

5 Funding, Timing, and Mechanism Rules

5 public rules should appear before drafting. According to NIH, PAR-25-357 is an R21 clinical trial not allowed opportunity. Grantsights records the live program deadline as October 16, 2027.

RuleSource-backed read
Funding numberPAR-25-357
MechanismR21 exploratory/developmental grant
Clinical trial statusNot allowed
Data typeExisting vision-related data
Live Grantsights deadlineOctober 16, 2027

According to NIH, applicants should explain how the data are of sufficient quality and already exist in analyzable form. According to USAspending.gov (opens in a new tab) data, the connected NEI assistance listing (93.867) had about $689.0M in FY2024 obligations across 2,876 awards active that year.

5 Route Choices Before Drafting

5 route choices matter because existing-data projects can belong in several places.

RouteBetter first read when...Why
PAR-25-357 R21Existing data can answer focused vision aimsR21 scope fits
Parent R21Vision fit is present but NEI route is not idealBroader route may fit
R01Dataset and aims justify larger multi-year scopeLarger mechanism may fit
Clinical-trial routeNew intervention assignment is plannedPAR-25-357 bars trials
Data-prep fundingData are not structured enoughAnalysis route is premature

The honest answer is that many secondary-data ideas are not ready because the data are not clean enough, the variables are not specific enough, or the team has not handled bias.

What NIH Doesn't Tell You

4 hidden readiness questions decide whether this belongs under PAR-25-357. NIH gives the public instructions, but it does not judge whether your dataset can carry the story until review.

First, NIH does not make data usable because they exist. Missing data, implausible values, selection bias, misclassification, and confounding can break the aims.

Second, NIH does not make all secondary analyses exploratory enough for R21. If the project is already a large, mature, multi-aim study, R01 may be better.

Third, NIH does not let an applicant avoid manuscript planning. The NOFO asks for milestones and timelines, including at least 2 manuscript submissions.

Fourth, NIH does not remove human-subjects responsibilities. Existing human data can still require records, privacy review, data-use agreements, and sharing plans.

Grantsights adds that missing layer with award-family context, data-to-aims checks, bad-fit signals, and a 21-day routing workflow.

According to NIH, applicants should describe existing datasets or databases and explain how the data will address the aims. According to NEI policy, clinical trials do not fit FOAs marked Clinical Trial Not Allowed. If the team needs new recruitment, new outcome collection, or a new intervention assignment, skip this route and choose a better route before writing a secondary-data proposal.

7 Evidence Assets That Make This R21 Stronger

7 evidence assets separate a strong NEI R21 from a dataset inventory. The strongest application reads like a planned analysis with clear outputs.

First, name the dataset and its source. Clinical trial data, epidemiologic data, clinical research data, EHR data, imaging data, registry data, and claims-linked resources may fit when vision outcomes are central.

Second, map each aim to variables. According to NIH, applicants should describe existing datasets or databases and explain how the data will address specific aims.

Third, describe data quality honestly. Include missingness, measurement timing, inclusion criteria, outcome definitions, and cleaning status.

Fourth, handle bias and confounding. According to NIH, applicants should include plans to control or account for potential biases and deficiencies in data.

Fifth, define manuscript milestones. One university in Pennsylvania with an NEI-funded clinical trial dataset, cleaned imaging variables, and a two-paper plan can give reviewers a concrete output path.

Sixth, show why the method matters. New methods can fit if the application explains how they improve future analysis of vision-related research data.

Seventh, decide whether the R21 is enough. If the analysis needs new data collection, new enrollment, or a large prospective design, avoid this route.

21-Day NEI R21 Routing Plan

21 days is enough to decide whether the team should draft. Days 1 through 3 should confirm source status, deadline, clinical-trial-not-allowed status, and NEI fit.

Days 4 through 7 should build the data inventory: dataset name, access status, variable list, quality checks, missingness, and data-use terms.

Days 8 through 12 should map aims to variables and test whether each outcome can be analyzed without new data collection.

Days 13 through 16 should write the bias and methods plan. This should include confounding, selection, missing data, and sensitivity analyses.

Days 17 through 21 should write the milestone plan, including at least 2 manuscript submissions and a realistic analysis schedule.

6 Reviewer-Risk Signals to Fix

6 risk signals should be fixed before submission. First, the application should not read like a fishing expedition.

Second, the data quality section should not be vague. Reviewers need proof that the data are ready.

Third, the project should not imply new data collection.

Fourth, the clinical-trial-not-allowed status should be respected.

Fifth, the publication plan should be specific.

Sixth, the application should explain why the work matters to NEI's mission, not just to general health data science.

7 Source-Backed Signals From PAR-25-357

7 signals from NIH should shape this page and the applicant's route memo. The first is that the project must be vision-related. A strong data science idea still needs a clear NEI mission fit.

The second is secondary-data scope. The data must already exist. If the project depends on new recruitment, new outcome collection, or a new intervention assignment, the route is likely wrong.

The third is clinical-trial-not-allowed status. Applicants should avoid any design that assigns an intervention to measure a health outcome. If there is doubt, the team should resolve it before the first draft.

The fourth signal is analyzable data. NIH asks applicants to describe datasets or databases and explain why the data are already structured, cleaned, and ready enough for the aims.

The fifth signal is bias handling. Missingness, measurement timing, selection, confounding, outcome definitions, and data linkage all need a plan.

The sixth signal is manuscript output. NIH asks for milestones and timelines, including at least 2 manuscript submissions. This gives the applicant a concrete way to prove the R21 is appropriately scoped.

The seventh signal is R21 scale. The project should be focused enough for 2 years and exploratory enough to justify the mechanism.

5 Signals To Check Before Drafting

5 signals show whether this R21 is ready to draft. Signal 1 is data access: the dataset must exist and be usable for the proposed aims.

Signal 2 is variable fit: each aim needs variables, timing, outcome definitions, and missingness checks.

Signal 3 is trial fit: no new intervention assignment can appear in a Clinical Trial Not Allowed R21.

Signal 4 is manuscript fit: NIH asks for milestones and timelines, including at least 2 manuscript submissions.

Signal 5 is program-staff contact: before drafting, write down the dataset-specific, bias-handling, manuscript-path, and R21-versus-R01 questions to ask the NEI contact listed in the notice.

5 Applicant Types and Best Next Move

5 applicant types commonly ask about PAR-25-357, and each should leave with a different route decision.

Applicant typeBest next moveWhy
Vision trial data teamBuild data-to-aims tableRoute may fit
EHR vision researcherCheck data quality and biasExisting data can fit
Team needing new recruitmentUse another routePrimary data work does not fit
Methods teamExplain future analysis valueMethod gain must be clear
Sponsored-project officeCheck trial status and DUA timingTechnical issues can block

The best answer is: use PAR-25-357 when the project uses existing vision-related data, is not a clinical trial, has analyzable data, and can produce clear manuscript outputs.

Sources

3 sources anchor this guide: the NIH NEI R21 NOFO, the federal listing, and the Grantsights public database record.

Next steps

Where to go next

These next pages help you move from research into a specific grant choice.

Frequently Asked Questions

What does PAR-25-357 fund?

NIH says the R21 supports vision-related secondary data analyses using existing data from clinical trials, epidemiologic studies, clinical research, or health care data resources.

Are clinical trials allowed?

No. PAR-25-357 is clinical trial not allowed.

Do data need to exist already?

Yes. NIH tells applicants to describe existing datasets or databases and show that the data are already structured, cleaned, and analyzable enough for the aims.

Are manuscript milestones required?

NIH asks applicants to include milestones and timelines with at least two manuscript submissions.

Who should skip this route?

Skip it if the project needs new primary data collection, a clinical trial, or data that are not yet usable.

Last updated October 10, 2026. Published by Grantsights from public sources. See our editorial policy, how we build and check guides, corrections and how we make money.