Skip to content

RFA-DA-26-056 Existing Data R21 Eligibility Guide

Updated 17 official sources

Quick Answer: RFA-DA-26-056 is NIDA's 2-year existing-data R21: up to $275,000 in direct costs, no new data collection, no clinical trials, and a next due date of December 3, 2026.

Full answer and how we checked it

Eligibility turns on 6 checks: dataset access, substance-use centrality, no primary data, no clinical trial, R21-sized aims, and an R21-versus-R01 route decision.

Who this is for: Substance-use researchers, HIV and prevention scientists, epidemiologists, data-science teams, neuroimaging researchers, health-services researchers, and sponsored-research offices deciding whether an existing-data R21 fits.

Data note: This guide uses NIH RFA-DA-26-056 (opens in a new tab), companion NIH RFA-DA-26-055 R01 (opens in a new tab), the predecessor RFA-DA-24-037 (opens in a new tab), NIH R21 guidance (opens in a new tab), NIH clinical trials policy (opens in a new tab), Grants.gov listing 357545 (opens in a new tab), NIH notices, award records from the NIH RePORTER API (opens in a new tab), and USAspending.gov (opens in a new tab) assistance-listing data, checked on October 18, 2026.

In this guide

RFA-DA-26-056 is not a general NIDA R21. According to NIH RFA-DA-26-056 (opens in a new tab), the route encourages analyses of public-use and other existing datasets to increase knowledge of substance-use etiology, trajectories, prevention, treatment, services, morbidity, mortality, and HIV-related outcomes.

Strong data teams can lose time when they treat "we have data" as the whole argument. Existing data only help when the application shows dataset access, variables, bias risks, and a question sized for 2 years.

Start from the live NIDA Existing Data R21 program record. Then compare NIH grants complete guide, NIH R01 R21 R03 K awards comparison, NIH Substance Use Existing Data R01 guide, what is a NOFO, and the Grantsights grant database.

October 2026 Source Check

As of October 18, 2026, RFA-DA-26-056 remains open and hasn't been reissued. Grantsights checked the RFA's key dates, budget, eligibility, and review sections on grants.nih.gov (opens in a new tab), matched the federal listing on Grants.gov (opens in a new tab), and compared the funded projects against NIH RePORTER award records.

The next due date is December 3, 2026, a standard date for new, renewal, resubmission, and revision applications. HIV/AIDS projects use that same date: NIH is removing separate AIDS due dates for January 2027 and later council rounds (NOT-OD-26-029 (opens in a new tab)), and the RFA's AIDS column reads "Not Applicable" from July 17, 2026 on. A December 3, 2026 application goes to scientific review in March 2027 and advisory council in May 2027, with an earliest start date of July 2027. Later due dates are July 15, 2027 and December 3, 2027, and the RFA expires December 4, 2027.

The source facts that control fit are narrow: primary data collection is not allowed, clinical trials are not allowed, the R21 budget is capped at $275,000 in direct costs over 2 years with no more than $200,000 in any single year, and the companion RFA-DA-26-055 R01 is the better comparison when the existing-data project is too large for a focused R21. According to NIH, NIDA intends to commit $2 million in FY 2026 for 2 to 4 awards, plus another $2 million for 2 to 4 HIV-focused awards, shared across this R21 and the R01.

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.

Grantsights Analysis: 13 Existing-Data R21 Awards Since FY2022

Grantsights analysis of NIH RePORTER records found 13 new NIDA R21 awards under RFA-DA-26-056 and its two R21 predecessors from FY2022 through FY2026. Method: on October 18, 2026, Grantsights queried the NIH RePORTER API (opens in a new tab) for every project record whose funding opportunity number is RFA-DA-26-056, RFA-DA-24-037, or RFA-DA-22-038 (paging in blocks of 500 with an offset; the full pull was 20 records), kept NIDA R21 records, and counted new (type 1) awards by fiscal year. Median costs use RePORTER's total-cost field.

Fiscal yearNew R21 awardsFunding opportunity
FY20223RFA-DA-22-038
FY20231RFA-DA-24-037
FY20244RFA-DA-24-037
FY20251RFA-DA-24-037
FY20264RFA-DA-26-056

Three patterns matter for budgeting. First, the 9 new awards from FY2022 to FY2025 had a median first-year total cost of $246,000. Five of those 9 show exactly $150,000 in year-one direct costs and $125,000 in year two, which is the $275,000 two-year cap split unevenly toward year one. Second, the 4 FY2026 awards look different: their single RePORTER records range from $207,585 to $476,808 (median $433,463), each project period runs into 2028, and none has a second-year record yet. That pattern fits NIH funding both budget years at once, but RePORTER doesn't label it, so don't plan around it. Third, the median total cost recorded so far per project, summing every fiscal-year record, is $434,500 across all 13.

Two FY2026 awards show more than $275,000 in RePORTER's direct-cost field. That isn't a cap waiver. NIH's consortium F&A policy (NOT-OD-05-004) (opens in a new tab) excludes subaward indirect costs when checking a direct-cost limit, so a team with a data partner on a subaward can carry those costs on top of the cap.

The award count also tells you something about odds. According to NIH, NIDA planned 4 to 8 awards across this R21 and the R01 in FY2026. RePORTER shows 4 new R21s and 9 new R01s under the two RFAs in FY2026, so NIDA funded more than the RFA's planning number, and the R01 drew more awards than the R21.

Which datasets funded R21s used

The RFA names the Adolescent Brain Cognitive Development (ABCD) Study and the Population Assessment of Tobacco and Health (PATH) Study, so Grantsights searched each project's title and abstract for 19 named data sources, matching the acronyms ABCD, PATH, and HEAL case-sensitively and every other name case-insensitively (for example "National Survey on Drug Use and Health," "Monitoring the Future," "NAHDAP," and "Add Health").

Data source named in the abstractProjects (of 13)Example award
ABCD Study4R21DA058314, University of California, San Diego
Health-system, military, or state records4R21DA065889, Massachusetts General Hospital
Investigator-held study data2R21DA057580, Virginia Tech
PATH Study1R21DA066670, University of Nebraska Medical Center
GWAS summary statistics1R21DA061372, Yale University
State policy and law data1R21DA060486, RAND Corporation
NSDUH, Monitoring the Future, NAHDAP, HEAL, NIDA Data Share0None by keyword

The rows other than ABCD and PATH come from reading the abstracts, since no shared keyword names them. The shift is the useful signal: all 4 ABCD projects were funded in FY2022 to FY2024, and none of the 4 FY2026 awards uses ABCD. Three of the FY2026 awards analyze linked records (homeless-clinic and Massachusetts death records, Military Health System prescriptions, and Maryland administrative and geospatial data), and the fourth harmonizes PATH Waves 7 to 9 with restricted biomarker data. For example, Massachusetts General Hospital received $476,808 in FY2026 for R21DA065889, which models overdose death among 71,914 adults seen at Boston Health Care for the Homeless Program from 2003 to 2022, according to its RePORTER abstract.

None of this means NSDUH, NAHDAP, or HEAL data can't win. It means a reviewer will judge the data-to-aims fit, not the brand of the dataset.

6 RFA-DA-26-056 Eligibility Stop Checks

6 stop checks sort a secondary-analysis idea into this R21 or out of it. The data must already exist, substance use must be the point of the question, nothing new can be collected, and the analysis has to finish inside a 2-year exploratory budget.

Stop checkStrong R21 signalRed flag
Data statusDataset already exists and access is clearNew enrollment or collection is needed
Topic fitSubstance use, SUD, HIV, prevention, treatment, or services is centralSubstance use is incidental
Trial statusNo NIH-defined clinical trialNew intervention assignment is planned
R21 scaleFocused exploratory analysisFull multi-year R01 is hidden inside
Bias planMissingness and confounding are addressedFishing expedition
Output pathFindings can guide prevention, services, policy, or future studyNo practical interpretation

Don't use this R21 to fund data cleaning as the main project. NIH says awards aren't meant for ongoing data analysis, for maintaining or distributing datasets, or for analyzing banked biospecimens that don't already have data, so the data should be ready enough to answer the aims.

5 Funding, Timing, and Mechanism Rules

5 public rules from the RFA set the frame for an existing-data R21 budget and timeline. According to NIH, RFA-DA-26-056 is an R21 with clinical trials not allowed, and any application that includes primary data collection won't be reviewed.

RuleSource-backed read
Funding numberRFA-DA-26-056
MechanismR21 exploratory/developmental grant, 2 years maximum
Clinical trial statusNot allowed
BudgetDirect costs up to $275,000 over 2 years, no more than $200,000 in one year
Next due dateDecember 3, 2026 (then July 15, 2027 and December 3, 2027)
AIDS due datesNone; HIV/AIDS applications use the standard dates
Eligible applicantsUniversities, nonprofits, for-profits, governments, tribes, and foreign organizations

According to NIH, the companion R01 is RFA-DA-26-055. The RFA lists two assistance listings, 93.279 (NIDA) and 93.273 (NIAAA). According to USAspending.gov (opens in a new tab) data, assistance listing 93.279 (Drug Use and Addiction Research Programs) had about $1,489.1M in FY2024 grant obligations across 4,651 awards. That's the whole NIDA research portfolio, not this RFA's set-aside.

6 NIH-Wide Changes That Apply to a December 2026 R21

6 NIH policy changes from 2025 and 2026 apply to anyone submitting on December 3, 2026. None of them changes the RFA's eligibility rules, but each can change how the application is built or how the award decision is made.

  1. No paylines. According to the NIH Extramural Nexus, November 21, 2025 (opens in a new tab), NIH's Unified Funding Strategy took effect with the January 2026 council round: institutes no longer use paylines and instead weigh review scores with priorities, scientific opportunity, and available funds. For an RFA this was already close to true, since the RFA says applications compete only with others submitted to this NOFO. Read NIH paylines explained for what a score means now.
  2. Simplified review. Under NOT-OD-24-010 (opens in a new tab), reviewers score Factor 1 (Importance of the Research) and Factor 2 (Rigor and Feasibility) from 1 to 9, and rate Factor 3 (Expertise and Resources) as appropriate or not. RFA-DA-26-056 uses this framework.
  3. Indirect costs. On January 5, 2026, the First Circuit affirmed the ruling that blocked NIH's 15% indirect cost cap in Massachusetts v. NIH (APLU summary (opens in a new tab)). Budget indirect costs at your negotiated rate; see the indirect cost rate negotiation guide.
  4. Common Forms. NOT-OD-26-018 (opens in a new tab) moved NIH to the Common Forms for the Biographical Sketch and Current and Pending (Other) Support, and NOT-OD-26-079 (opens in a new tab) made them required for due dates on or after May 8, 2026. The old NIH biosketch format now stops a submission with an error. The NIH biosketch guide covers the SciENcv steps.
  5. Six-application limit. According to NOT-OD-25-132 (opens in a new tab), NIH accepts no more than 6 new, renewal, resubmission, or revision applications per PI per calendar year, and it won't treat applications substantially developed by AI as the applicant's original ideas. A lab sending this R21 and the companion R01 in the same year uses 2 of those 6.
  6. Continuing resolution. According to NOT-OD-26-131 (opens in a new tab), NIH is operating under a continuing resolution through December 11, 2026 at the FY2026 level. That affects FY2027 money, which is when a December 2026 application would be paid.

5 R21 Versus R01 Route Choices

5 route choices matter because R21 and R01 are often confused for existing-data work. The RePORTER counts above show the R01 drew 9 of the 13 new FY2026 awards under the two RFAs.

RouteBetter first read when...Why
RFA-DA-26-056 R21Dataset can answer focused exploratory aimsR21 scope fits
RFA-DA-26-055 R01A larger existing-data study is justifiedR01 scope fits
Parent R21NIDA RFA fit is weak but existing-data idea is earlyBroader route may fit
Parent R01Topic is mature and multi-yearLarger mechanism may fit
Data-prep routeData are not readyAnalysis route is premature

The best first move is a data-to-aims table. If every aim can't be tied to existing variables, access rights, and bias controls, the project isn't ready. For the general mechanism, see the NIH parent R21 guide; for the R01 side, the existing-data R01 guide covers its 5-year scope and budget questions.

What NIH Doesn't Tell You

4 hidden readiness questions decide whether RFA-DA-26-056 is worth pursuing.

First, NIH does not make a dataset strong because it is large. Large data can create false confidence if missingness, measurement timing, subgroup power, and confounding are weak. The RFA asks applicants to address the risk of spurious findings from multiple analyses and to report effect sizes, not just p-values.

Second, NIH does not permit primary data collection here. A project that depends on new surveys, new interviews, focus groups, or new assays on stored samples needs a different route.

Third, NIH does not make all secondary analysis novel. The RFA says the proposed research must be distinct from the aims and methods of the primary study that produced the data, and from other funded work using the same data. With ABCD, that's a real hurdle, because any approved investigator can download the same release.

Fourth, NIH does not make R21 a fallback for underbuilt R01s. The R21 should have a focused question and a clear 2-year output.

There's also a requirement that's easy to miss. Applicants using ABCD data, "and others where appropriate," must include a plan to pre-register their statistical analyses with a public registry or as a registered report. A common mistake is to treat pre-registration as optional because it isn't in the page limits table.

Finally, peer review isn't handled by a standing CSR study section. NIH says a Scientific Review Group convened by NIDA reviews these applications, the National Advisory Council on Drug Abuse does second-level review, and appeals of initial peer review aren't accepted for this NOFO.

7 Evidence Assets That Make This R21 Stronger

7 evidence assets help the application read as more than a dataset inventory.

First, name the dataset and access status. ABCD, PATH, NAHDAP, NIDA Data Share, the Human Connectome Project, state and local data, administrative records, and HIV cohorts such as MWCCS are all named in the RFA as fitting sources when the project has real access.

Second, map variables to aims. The application should show which exposure, outcome, covariate, timepoint, and subgroup variables answer each aim.

Third, handle bias and false discovery. According to NIH, applicants should consider spurious findings when using large data sets or multiple analyses.

Fourth, show substance-use centrality. A project can include mental health, family, services, HIV, or neurodevelopment, but substance use shouldn't be an afterthought.

Fifth, define the R21 output. For example, one university in California, the University of California, San Diego, received $237,000 in FY2023 for R21DA058314 to build and preliminarily validate screening tools for early high-risk drinking and drug use from ABCD data, according to its RePORTER record. That's a clean R21 output: one tool, one validation step, one data source.

Sixth, show translation path. NIH asks applicants to be explicit about ways findings can inform prevention, services, policy, or future intervention work when possible.

Seventh, compare the R01. If the dataset can support a full multi-year program with several mature aims, RFA-DA-26-055 may be a better fit. The specific aims guide helps test whether 2 aims fit in 2 years.

7 Source-Backed Search Questions

7 questions come up most often about this existing-data R21. The first is whether the dataset must already exist. Yes. NIH says primary data collection is not allowed, and that includes new assays on banked biospecimens.

The second is whether an emulated trial counts as a clinical trial. It doesn't, as long as no one is prospectively assigned to an intervention; the RFA even lists emulated trials from observational cohorts as an HIV research interest. A team planning a new intervention assignment needs another route.

The third is whether qualitative data can fit. NIH describes public-use and other extant community-based or clinical datasets and notes that existing data may include qualitative information and mixed data.

The fourth is whether NSDUH or HEAL data can fit. The RFA doesn't name either one, but it welcomes public-use, administrative, and other NIH-supported data. Confirm with the scientific contact, and plan for any restricted-file approval before the due date.

The fifth is whether foreign organizations can apply. NIH lists foreign organizations and foreign components as eligible for this RFA, so applicants should still verify registration and institutional rules, including the NCAGE code a foreign entity needs for SAM registration.

The sixth is whether the R21 should lead to an R01. Sometimes. The honest read is that the R21 should produce a focused analysis or pilot signal, while RFA-DA-26-055 is better when the project is already larger.

The seventh is whether a Data Management and Sharing Plan is needed when no new data are collected. Yes. The RFA says every application must address one, regardless of budget. Derived variables, code, and linked files count as scientific data, and the data management plan guide covers the format.

21-Day R21 Routing Plan

21 days before committing to a December 3, 2026 draft is enough for a go or no-go call on this R21. Days 1 through 3 should confirm the RFA is still open, the due date, trial status, the $275,000 cap, and the companion R01 comparison.

Days 4 through 7 should build the dataset inventory: source, access, data-use terms, variables, missingness, dates, and population. Restricted files such as PATH biomarker data need their own approval timeline, and teams often underestimate it.

Days 8 through 12 should link each aim to named variables and waves in the dataset and confirm none of them needs new collection. If ABCD is involved, draft the pre-registration plan now.

Days 13 through 16 should write the bias plan: confounding, selection, missing data, subgroup limits, multiple testing, effect sizes, and sensitivity analyses.

Days 17 through 21 should decide R21 versus R01, count the PI's applications against the 6-per-year limit, generate Common Forms in SciENcv, and prepare the program-staff question memo.

6 Reviewer-Risk Signals to Fix

6 reviewer concerns show up often enough in existing-data applications to fix before December 3. First, the application shouldn't imply new data collection. A red flag is any mention of "pilot interviews" or "follow-up surveys" in the approach.

Second, the dataset access plan should be specific: who holds the data, what agreement covers it, and when access starts.

Third, the variables shouldn't be vague. Reviewers need to see that the data can answer the aims under Factor 2, Rigor and Feasibility.

Fourth, the project shouldn't read like a fishing expedition. An application fails because the analysis plan lists outcomes without a correction strategy or a stated primary hypothesis.

Fifth, the work shouldn't duplicate the primary study's original aims or another funded project on the same data. Search RePORTER for the dataset name before writing.

Sixth, the R21 should state what the 2-year analysis will produce and how that output informs the next step. If the application is scored but not funded, the summary statement guide explains how to read the critique before a resubmission.

5 Signals To Check Before Drafting

5 readiness signals separate a draftable existing-data R21 from one that needs another cycle. Signal 1 is data access: the dataset must already exist, and any data use agreement or restricted-file approval should be signed or scheduled.

Signal 2 is rule fit: primary data collection and clinical trials are not allowed.

Signal 3 is mechanism fit: the project must be small enough for an R21 and not a hidden R01.

Signal 4 is review risk: the team needs a variable map, missingness plan, confounding plan, and false-discovery plan.

Signal 5 is a short note to NIDA: the RFA's scientific contact in Section VII is Marsha F. Lopez, Ph.D., M.H.S., and the useful questions are whether the named dataset fits and whether the scope reads as R21 or R01.

5 Applicant Types and Best Next Move

5 applicant types commonly ask about RFA-DA-26-056.

Applicant typeBest next moveWhy
Existing-data PIBuild data-to-aims tableRoute may fit
Team needing new survey dataUse another routePrimary data is not allowed
Mature data-science teamCompare RFA-DA-26-055R01 may fit better
HIV and SUD services teamConfirm dataset and NIDA fitTopic fit matters
Sponsored-project officeCheck budget, subaward F&A, and trial statusTechnical fit can block

The best answer is: use RFA-DA-26-056 when the data already exist, the question is focused, no clinical trial is proposed, and the R21 can produce a clear analysis output.

If the dataset isn't accessible, the variables don't map to aims, or the question is too large for 2 years, don't use this R21 yet. For other addiction funding, including service grants, see grants for substance abuse programs or search the Grantsights grant database.

Sources

6 sources anchor this guide: the NIH R21 NOFO, the companion R01 NOFO, the federal listing, NIH policy notices, NIH RePORTER award records, 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 RFA-DA-26-056 fund?

NIH says the R21 supports analyses of public-use and other existing datasets to answer substance-use, HIV, prevention, treatment, services, and related research questions.

Can applicants collect new primary data?

No. NIH states that primary data collection is not allowed, and it lists interviews, surveys, focus groups, expert panel feedback, and analysis of banked biospecimens as nonresponsive.

Are clinical trials allowed?

No. RFA-DA-26-056 is clinical trial not allowed.

What is the R21 budget?

NIH states that direct costs may not exceed $275,000 over 2 years, with no more than $200,000 in any single year.

What is the companion R01?

RFA-DA-26-055 is the companion R01 for larger existing-data substance-use projects.

When are the remaining due dates?

December 3, 2026, July 15, 2027, and December 3, 2027. The RFA expires December 4, 2027, and HIV/AIDS applications now use the same standard dates.

How many awards has this R21 series made?

NIH RePORTER lists 13 new NIDA R21 awards from FY2022 to FY2026 under RFA-DA-26-056 and its two predecessors, including 4 in FY2026.

Last updated October 18, 2026. Corrected October 18, 2026: what changed. Published by Grantsights from public sources. See our editorial policy, how we build and check guides, corrections and how we make money.