Non-R01 Individual Investigator Mechanisms: The Growth of R21 Applications

Aug 04 2014 Published by under Uncategorized

In a previous post about R01s, I noted that the fraction of the NIH budget going to R01s decreased over the period from FY2003 to FY2013. This fact is of concern, of course, for a variety of reasons. But first, it is important to understand the observation as completely as we can. One set of factors that has contributed is the growth in the number of R21 awards, driven in large part by a huge increase in the number of R21 applications and the introduction of additional mechanisms.

For the R21 mechanism, the numbers of applications and awards NIH-wide for the period from FY2003 to FY2013 are shown below:

R21 Apps Awards plot

These data reveal that the number of applications increased more than 2.6-fold from 5283 in FY2003 to almost 14000 in FY2012. NIH responded to this "proposal pressure" by increasing the number of awards from 1255 in FY2003 to almost 2000 in FY2012. Nonetheless, the success rate for R21s has remained between 12.9 and 14.9 % over the past five years, generally 2 or more percentile points below the R01 success rate, even for new (as opposed to competing renewal) applications. Thus, the competition for R21 awards is more severe that it is for R01 awards, a fact which is not, in my experience, widely appreciated by the all in the community.

When I was Director of NIGMS, we decided to stop accepting unsolicited R21 applications. This decision was made for two reasons. First, we found the peer review process for R21s quite frustrating. NIGMS was trying to use the R21 mechanism to support "high risk-high potential reward" research, that is, new ideas for which a modest investment could provide a proof of principle that could be used to drive future inquiry. However, despite efforts by CSR and NIGMS staff to orient reviewers, we frequently received scores and summary statements that did not align e.g. 'This is a potentially important and impactful project, but there is no preliminary data' and a bad score or 'This is a solid proposal supported by much preliminary data' with a good score. Because of this, we often struggled to develop sensible paylists. NIH made this problem worse by using the R21 mechanism for many other purposes other than the "high risk-high potential reward" goal. This confused reviewers, applicants, and even NIH staff.

Second, we had misgivings about whether the duration of the R21 and the size of the award would, in general, support substantial research compared to taking the same funds and supporting a smaller number of R01-sized grants. This led to the EUREKA R01 awards, used by NIGMS and a few other institutes.

I do not know of any studies that bear of the success of R21 awards in promoting scientific discovery or in keeping investigators "in the game". One interesting observation is that, in FY2013, more than 2300 (62%) of the R21 awards were held by investigators who had no other R21 or R01 awards (competing or non-competing) in the same year. Looking back to earlier year, one could track subsequent results for such investigators if that would be of interest.

12 responses so far

  • AnotherAnon says:

    Thus, the competition for R21 awards is more severe that it is for R01 awards, a fact which is not, in my experience, widely appreciated by the all in the community

    Might this vary by institute? An NIH person just tried to sell me on the opposite. The idea being that R21 applications have more junior investigators than the R01 applicant pool. So for a junior person who is pre-R01 the R21 (or 03) is a better starting point.

    • datahound says:

      Good question. There are some differences in success rate, most notably, NIAID which has a higher success rate and also funds the most R21 awards. Note that these data include both awards from the "parent" R21 announcement and more specialized announcement.

      Here are success rate data by IC (along with the number of awards):

      IC R21 R01 (New) Number of New R21 Awards

      NIAID 17% 12% 346
      NCI 11% 13% 284
      NINDS 17% 17% 207
      NIMH 17% 17% 130
      NIDA 19% 18% 106
      NIA 11% 11% 102
      NICHD 8% 11% 99
      NHLBI 14% 14% 77
      NIBIB 11% 15% 74
      NIAMS 13% 15% 66
      NIAAA 17% 19% 59
      NEI 14% 23% 51
      NIEHS 11% 12% 50
      NIDDK 9% 15% 28
      NIGMS 9% 15% 23
      NIDCD 12% 23% 21
      NIDCR 10% 19% 19
      NHGRI 13% 24% 13
      NINR 5% 11% 9
      NCCAM 6% 12% 8
      FIC 36% 19% 4
      NIMHD 13% 4% 1
      NLM 3% 13% 1

  • odyssey says:

    Interesting. Supports what I've thought about R21's in general. And my experience with reviewers ("not enough preliminary data!"). Do the same issues carry over the R03's? I know there is some general feeling that R03's are too small to be particularly useful, but my feeling is, if they were awarded and used in the manner the were designed to be, they could be quite useful.

  • AcademicLurker says:

    Once the length for the RO1 proposal was reduced from 25 pages to 12, the rational for applying for R21's largely disappeared. With the R21 you have the same amount of work, a similar or even lower chance of success, and less $$ in the event that you are actually funded.

  • anonymous postdoc says:

    Boy howdy would I love to see the sex distribution of R21s vis a vis R01. Given that the effort expended is similar, as AcademicLurker notes, but the reward is so much less.

    Furthermore, I worry that the idea that the R21 is meant to fund "risky" science likely translates into reviewer brains as funding "risky" scientists (read: junior, often female), as NIH seems already to be operating on a de facto, if not de jure, "people, not projects" funding strategy.

  • Joe says:

    The last R21/R03 panel I served on, we were told that the R21 for NIAID did not need to be high-risk. So, even though some reviewers used "high risk - high reward" as a selling point for their favorite applications, applications that were not in that mold, could not be dinged on that account. On the prelim data front, we are reminded every time that prelim data is not necessary for an R21 and applications cannot be criticized for not having it. Of course, those that have nice prelim data are highly praised for it. I wouldn't send in an R21 application without it.

    I've seen a good number of R21 applications that were cut-down versions of failed R01 applications. Sometimes the budget retains traces of the 5yr R01 budget, sometimes the applicant says as much in the biosketch personal statement. So I get the feeling that people are using the R21 just to hang on and keep the lab alive until full funding can be obtained. So I would be very interested in seeing an analysis of what happens to investigators that at some point only have R21 funding.

  • drugmonkey says:

    The R21 is both Exploratory and *Developmental*. Acting like this is a problem just because on IC only wants the Exploratory (high risk/reward) is misplaced.

    I think the failing for PO/study section communication lies with Program, personally, for not being clearer on their own use of the mechanism and for refusing to guide review more clearly.

    • datahound says:

      NIGMS stated its intention of using the R21 for "High Risk-High Impact" research going back to at least 1997 (see and Through program announcements and some other venues, NIGMS staff tried to be clear that this was the intent. However, as I indicated, NIH (including NIGMS) bears it share of blame for using the same mechanism in many different ways, leading to confusion among both applicants and reviewers.

      I know from my time on study section, at least some NIGMS staff did try to guide discussion of R21s to align with NIGMS' programmatic goals for the mechanism, but this clearly was not fully successful. This is due, in part at least, to the fact that critiques are prepared prior to the meeting.

  • k says:

    I spoke with an NCI PO this morning about my 11% R21 that will not be funded, and was told that the culture is turning and that small RO1s disguised as an R21 no longer have a chance at this Institute.

    • datahound says:

      Frustrating. Again, different ICs use this mechanism in different ways and NCI is often different from the other ICs. For example, NCI has its own parent "omnibus" R21 announcement (, different from the NIH-wide one. This states

      "Applications for R21 awards should describe projects distinct from those supported through the traditional R01 mechanism. For example, such projects could assess the feasibility of a novel area of investigation or a new experimental system that has the potential to enhance health-related research. Another example could include the unique and innovative use of an existing methodology to explore a new scientific area. Conversely, long-term projects, or projects designed to increase knowledge in a well-established area, will not be considered for R21 awards."

  • […] a recent post, I highlighted the growth in the number of applications and, to a lesser extent, awards for R21s. […]

Leave a Reply