NIH Grant Proposal Example: Read Samples Like a Reviewer

Read in four passes. Do not begin by highlighting polished sentences.

In our editorial work, the most useful information gain comes from pairing an application with its summary statement. A polished aim shows what the applicant wrote; the review shows which dependency, evidence, or risk a reader actually noticed. We created the four-pass map below to preserve that causal trail instead of turning funded prose into a generic writing template.

For every reviewer statement, record the criterion, quoted judgment in your private notes, likely source section, supporting passage location, and what the applicant did or failed to do. This is more useful than copying a Specific Aims layout because it reveals how the application generated confidence.

The example below is invented. It demonstrates reasoning and contains no language from a funded application.

Overall objective: Determine whether a low-cost home monitoring protocol can identify early deterioration after discharge for adults with chronic respiratory disease.

Premise: Readmissions often follow a measurable change in symptoms and oxygenation, but evidence is insufficient on which change threshold transfers across routine-care settings.

Aim 1: Derive a prespecified risk rule in two health systems using daily symptoms, oxygen saturation, and prior utilization.

Aim 2: Externally validate calibration and decision-curve utility in a third system, including prespecified subgroup uncertainty.

Expected outcome: A transportability-tested risk rule and an explicit account of where home monitoring is not reliable. The study will not establish that acting on the rule improves clinical outcomes.

This example exposes its dependency: Aim 2 depends on a stable definition and feasible data collection from Aim 1, but the validation dataset must remain independent. The outcome is bounded to prediction and decision utility. It does not claim intervention effectiveness.

For each aim, answer five questions:

1. What decision-relevant uncertainty does the aim remove?
2. What observable result counts as success?
3. What evidence makes the approach feasible now?
4. What happens if the expected result is wrong?
5. Does failure of an earlier aim make the later aim impossible?

The serial-collapse pattern appears when Aim 2 cannot begin unless Aim 1 produces the preferred result. Repair it with an independent dataset, alternative branch, or a later aim that remains informative under multiple outcomes.

The technique-list pattern appears when aims are methods rather than scientific decisions. Repair it by naming the uncertainty each method resolves.

The guaranteed-outcome pattern appears when the proposal treats a hypothesis as a result. Repair it by describing what will be learned under supportive, null, and conflicting outcomes.

Use our walkthrough to strengthen the Specific Aims argument, but let the current funding opportunity determine the final components and limits.

NIH's review framework and the funding opportunity define what reviewers are asked to judge. Do not paste criterion labels into headings and assume alignment. Translate each criterion into a question the application must answer.

The map must be opportunity-specific. A sample can show how one team handled feasibility; it cannot establish that the same evidence satisfies your reviewers.

A funded application is not a perfect application. Its summary statement may reveal weaknesses reviewers tolerated because the overall case remained compelling. That tension is valuable.

Separate three categories:

• portable reasoning: clear aim logic, bounded inference, visible contingency;
• project-specific advantage: unique cohort, tool, expertise, or institutional access;
• historical compliance: old forms, page layouts, biosketch rules, data policies, and criterion wording.

Only the first category is directly reusable. The second must be earned with your evidence. The third must be replaced with current instructions.

Check whether each reviewer concern has a proposal location after mentors return comments.

Use at least three official samples when available. One sample can make an idiosyncratic choice look universal.

Replace bracketed fields in your working copy. Do not place confidential proposal text in an unapproved tool or shared file.

• Verify the current opportunity number, due date, activity code, and application guide.
• Label every sample with its submission year and mechanism.
• Map each aim to a decision, outcome, evidence burden, risk, and alternative.
• Confirm later aims remain informative if earlier findings are null.
• Trace significance claims to current literature and preliminary claims to owned evidence.
• Check that team roles and environment support named tasks.
• Reconcile human-subjects, animal, data-management, authentication, and other applicable plans.
• Remove copied phrasing and obsolete form assumptions.
• Ask a reviewer unfamiliar with the project to state the premise, aims, and bounded outcome.

An NIH grant proposal example is most useful when it makes reviewer logic visible. It becomes dangerous when a historical application is mistaken for today's instructions or another project's language is mistaken for your evidence.

Sources accessed September 11, 2026.

1. Sample Applications and Documents, NIH Grants & Funding.
2. How to Apply: Application Guide, NIH Grants & Funding.
3. Simplifying the Framework for NIH Peer Review, NIH.
4. Sample Grant Applications, NIDCD.