NIH NRSA Fellowships: Timeline, tips, statistics

Ruth L. Kirschstein National Research Service Awards (NRSAs) are a popular group of fellowship awards offered by the National Institutes of Health (NIH). The majority of the funding goes to so-called institutional NRSAs, which include the Medical Scientist Training Program (MSTP) MD/PhD training mechanism, as well as a variety of smaller and more targeted fellowships offered by individual universities. For example, one institutional NRSA at UCLA is the Neuroimaging Training Program. There are also individual NRSA awards, which are applied for by the trainee, themselves. These individual awards are further subdivided by the career stage of the applicant:

  • F30 - Predoctoral trainees in dual-degree programs
  • F31 - Predoctoral trainees
  • F32 - Postdocs
  • F33 - Senior fellows

My Experience

Although being in the MSTP program means that someone has promised to fund you for each part of your graduate training, many students are still interested in applying for an individual NRSA on their own. There are various reasons for doing so, including that it is good experience for future “real” grant writing, and is also an accomplishment for the CV. The biggest reason that pushed me to try for one, though, is that it provided a nice structure and deadline to force me to consolidate my own thoughts, hypotheses, and goals as I prepared to pitch my dissertation project proposal to my PhD committee. I figured, since I had to go through many of the NRSA steps anyway for my oral qualifying exam, why not go a bit further and turn my proposal into a full NRSA grant application?

Tips

  • Begin early. The process takes 1-3 months to prepare your application, depending on how leisurely you take it, and then a full 6-8 months from application submission to funding decision. If you don't get funded on the first try, then it can take another 6 months for the next cycle (or possibly even 4-8 more on top of that if you missed the next submission deadline(s) because you were waiting for your first application).
  • Read the goals of NRSA program announcement carefully, and make sure that your research and training proposal are directly in the spirit of these goals.
  • Similarly, read through the review criteria section of the program announcement carefully, and make sure that your application makes the answers to these questions clear for the reviewers.
  • Send in a supplemental information update. This is optional, but the comments in my summary statement made it clear that the reviewers liked to see that I was making progress. The deadline is typically 1 month before your SRG meeting date.

Timeline

Date Action
8/2/10 Submitted to UCLA Graduate Division
8/3/10 eRA Commons update: Application entered into system
8/6/10 Application assigned to NIAAA on eRA Commons
8/6/10 eRA Commons update: Scientific Review Group review pending. Refer any questions to the Scientific Review Administrator
9/22/10 Submitted supplementary material for review
11/2/10 Scientific Review Group (SRG) meeting
11/5/10 Priority score posted to eRA Commons
11/5/10 eRA Commons update: Scientific Review Group review completed. Refer any questions to Program Official
11/22/10 Summary statement posted to eRA Commons
12/27/10 Just-in-time (JIT) request for additional information
2/16/11 Advisory Council (AC) meeting
4/5/11 Notice of Award posted on eRA Commons

Links

  • F-Kiosk: Links to the most current program announcements for the F30 and F31 NRSAs.
  • SF424 (R&R) Forms: Contains the most current application guide (be sure to choose the “fellowship” one), as well as the supplemental templates for biosketches, letters of references, etc.
  • NIH F30 NRSA Program Announcement PA-10-107: Information on submitting an application for this specific program announcement. It supersedes any information in the more general SF424 guide. These PA pages are constantly getting replaced or amended, so be sure to check back several times during the cycle in case of updates.
  • eRA Commons: NIH tracking system for submitted/funded grants.
  • UCLA Graduate Division NRSA page: Institution-specific information about completing and submitting NRSA applications. The school's deadline will come before the NIH deadline, so plan ahead.

NRSA Statistics

When I was working on my NRSA, I was pretty curious about what my chances were of actually getting funded. Did this vary by institute? Is it getting more difficult as time goes by? What sorts of projects get funded? How do the answers to these questions differ for the F30 activity code for MD/PhD and other dual degree students?

NIH Data Book

The first resource I found was the NIH Data Book, which has a whole section about fellowship funding. Unfortunately, most of the data graphics there are kind of confusing because they use lots of stacked bar plots and focus on funding in terms of dollars (whereas I cared more about funding in terms of number of new projects). Plus, they didn't make it easy for me to compare F30s to F31s.

  • One exception is the F31 NRSA success rate plot, which is very valuable because the NIH does not release this information about unfunded projects anywhere else.

NIH RePORTER

At any rate, I kept looking for another option to explore these data. I knew about the NIH RePORTER tool, which is a way for anyone to look up information about individual NIH grants. Luckily, it has also been continually upgraded so that now entire batches of search results can be downloaded as .csv files.

  1. Select Activity Code → Fellowships (Fs) → F31 Predoctoral and All Other Fs
  2. For Fiscal Year (FY), select a batch like 2008-2010
  3. Hit query
  4. Select Export → All Projects, and then click the GO button.
    • Only 5000 projects can be exported at once, so if there are more than that in the search results, go back and select fewer years
  5. Select Format → CSV
  6. Select Columns to Export → All
    • But then uncheck the box for Project Abstracts
  7. Hit Select

Data Analysis

Using this procedure, you can retrieve batches for 2008-2010, 2005-2007, and 2001-2004. Together, this gives you access to detailed information concerning all of the predoctoral NRSAs that have been funded over the past ten years! I loaded this dataset into R and set about exploring the data.

10-year Summaries

Between fiscal years 2001 and 2010, UCLA, Johns Hopkins and UNC had the most new predoctoral NRSAs - averaging more than 15 per year. While the top 10 institutions account for 24% of the new projects over this period, there is a long tail of institutions who average several per year.

New projects, by institution (top 50)

The most common NIH institutes are NINDS and NIMH, which both focus on the brain, and together account for 37% of all the new NRSAs over this period. We shall see next that this is in line with most common departments as well.

New projects, by NIH institute

Psychology is, by far, the most common departmental affiliation. There are more than double the number of psychology projects than there are pharmacology projects, which is the next most common department.

New projects, by department


If we look at the total number of active NRSAs in each year of the last decade, we see a gradual upward trend for both F30 and F31 activities. An increasing proportion of NRSAs are F30s, which makes sense based on the expanding list of NIH institutes that support them. Towards the end of the decade, there appears to be a slowdown of new NRSAs (budget cuts?), although this effect may be hidden by the number of previously funded NRSAs that are still active.

Total active projects per year, by activity code

If instead we directly look at the number of new NRSAs per year, we see a dramatic drop in the number of newly-funded projects - with a drop by an entire third between 2009 and 2010 for both F30 and F31 activities. In 2010, it is also useful to see that F30s now make up approximately 20% of all new individual predoctoral NRSAs.

New projects per year, by activity code

Since the NIH does not release specific information on projects that weren't funded, we can't calculate the actual success rate. However, based off the suffixes that are added to project numbers upon resubmission, we can look at the fraction of funded projects that are funded on their first try (let's call this the Fraction on First Try, or FoFT). While this will be higher than the actual success rate (since some unfunded projects will not be resubmitted, or even if they are, will not be funded on any subsequent attempt), it is still possibly a sensitive indicator of relative “success”. Here we see that the FoFT has been steadily decreasing over the last decade for both F30 and F31 activities, with the success of F30s running slightly above that of F31s. A noteable exception to this trend comes in 2007-2008, where there is a boost in the success of F30s. We shall see where this comes from when we break these trends down further…

Fraction of new projects that were funded on the first try


Breakdown by Institution and Institute

If we breakdown the number of new NRSAs per year by the institution (here we're looking at the top 25 schools, ordered based on the 10-year summary numbers, above), we see some interesting features. First of all, there is a large amount of within-school variability. This makes sense based on the low numbers we're talking about here, but also based off my own experience that the publicity of NRSAs can vary from year to year depending on whether there are workshops/seminars available to facilitate the process. It is also interesting to look at the relative prominence of F30s vs. F31s at each school. At UCLA, the number of F31s far exceeds the number of F30s - perhaps due to the large graduate programs in psychology and neuroscience. However, at Washington University, where there is an exceptionally large MD/PhD program, the number of F30s can actually exceed the number of F31s.

  • While less informative and more confusing, this data can also be viewed as a spaghetti plot.
New projects per year, by institution

Now let's look at the breakdown by NIH institute. Here we see that some institutes are pretty stable in the number of NRSAs they fund (NIDCD), while others have large variations year-to-year (NINR). Some are on an upward trend (NINDS, NIAAA), while some have been on a downward trend (NIMH, NIGMS). This figure also reveals which institutes fund F30 awards (of course this info is also directly available at the NIH), and when some of them came on board. This brings us to one of the most interesting features in this view of the data: When a new institute opens up to F30s, there is a bonanza of funded projects, which can exceed the entire number of F31s funded by the institute, and can even exceed the number of F30s funded by the big players like NINDS. What this means to me is that if I was applying for an NRSA again, and NCI happened to open up to F30s in the same year, I would try hard to target my project to this institute if possible. Update: NCI is now accepting F30s in 2011! See PA-11-110

  • These plots are scaled within each institute, in order to better reveal trends. A plot using a constant scale is also available, and better shows the difference in numbers between institute: Constant scale plot
New projects per year, by NIH institute

The R code used to generate these figures is available here.

training/nrsa.txt · Last modified: 2011/04/10 12:01 pm PDT by John Colby
 
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Noncommercial-Share Alike 4.0 International
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki