NIH Says: "Fuck R21s and R03s"

Nov 01 2008 Published by under NIH

The NIH has finally realized what a total fucking boondoggle the R03 and R21 grant mechanisms are:

The NIH has found that the use of Small Grants (R03) and the NIH Exploratory/Developmental Research Grants (R21) has increased over the last few years6. However, recent analyses indicate that a smaller proportion of individuals with initial R21 or R03 grant support subsequently apply for and obtain R01-equivalent funding. In addition, the initial success rate for R21 applications often is lower than for R01 applications. Since R03 and R21 grants are limited in scope and period of support, they may not be the most effective way to launch an independent research career. Accordingly, the NIH encourages New Investigators, particularly ESIs, to apply for R01 grants when seeking first-time funding from the NIH.


For those that aren't already familiar, R03s are limited to two years and $100,000 total direct costs, R21s are limited to two years and $275,000 total direct costs, and neither allow for competitive renewal after the two years. This is in contrast to R01s, which are limited to five years and theoretically unlimited direct costs (although a maximal modular direct cost budget is $1,250,000), and which allow for unlimited numbers of competitive renewals. This is important, because competitive renewals have a *much* higher success rate than new applications.
Dumbfuck program officers have traditionally suggested that newly independent applicants "get their feet wet" with R03s and R21s, before applying for an R01. When I was starting my own lab, I called a relevant program officer, introduced myself, and asked for suggestions. The first words out of his motherfucking mouth were "R03 R21". That fucking god I had friends and colleagues to tell me how totally ridic that advice was. What a stupid fucking dipshit that program officer was.
Well, NIH has finally taken the explicit position that R03s and R21s are totally inappropriate for newly independent investigators, and that they are a TOTAL FUCKING DEAD END!!! They don't lead to R01 funding; they have the opposite effect, of preventing the investigator from achieving R01 funding.
Comrade PhysioProf takes the stronger position that R03s and R21s are totally fucking STUPID for *any* investigator to apply for. If you are going to the trouble to write a grant application, write a motherfucking R01! Think about this way: You're gonna write a ten page R03 or a 15 page R21, with the maximum payoff of 8% or 22% of a maximal modular R01 budget, respectively, and *no* future benefit of the *way* higher success rates for R01 competitive renewals, when you can just write a 25 page R01!?!? (Yeah, the page limits are going to change, but the point is the same.)
R21s and R03s--no matter how hard the dumbfuck SROs and study section chairs try to stop it--are *always* reviewed like "mini R01s". The NIH even explicitly recognizes that R21s have a *lower* success rate than R01s. You gotta have motherfucking rocks in your head to write an R21!
And one other thing: What the fuck is the deal with using the traditional study section peer review mechanism for piddly ass little chunks of change like R03s and R21s?? What a massive waste of reviewer and administrative time and effort to use study section panels to review these punky little turds. NIH program should make funding decisions on these things administratively using the same system as NSF uses for many of its grants. Two or three external written reviews, and program makes the decision whether to fund or not.

12 responses so far

  • Funny you should write this now because I just came from commenting on writedit's post with this:

    Good advice, both your longstanding position and now the NIH recommendation. The payoff-to-time expenditure ratio is in the long run much better for the R01 than a R03 or R21. R21s are especially a nightmare because they are still often reviewed together with R01s (which also leaves me a bit skeptical that clustering new R01 apps will provide any real advantage to the New Investigators.). I�m now seeing R03s pulled out into their own SEPs but 1) reviewers still have to be reminded they are not R01s and 2) it seems like an awful lot of effort, for both reviewers and applicants, for a $50K/yr, 2 yr award if they are still not practically being judged as high-risk pilot projects.

  • msphd says:

    Great post!
    My university seems to know this, since postdocs are not allowed to apply for R03s or R21s. They basically lump them in with R01s in terms of "institutional support."
    God forbid we should get any of that, what's it called?
    Oh yeah, "independence."

  • Very interesting... I work with a group of Canadian PIs who sometimes apply for NIH funding. We basically only have a chance if we respond to a specific RFA. We applied in the summer to an RFA that had only an R21 component, no R01 equivalent. Still waiting for reviews. I hope it wasn't a complete waste of time.

  • DrugMonkey says:

    I hope it wasn't a complete waste of time.
    Some of us have a more nuanced view of the value of R21s and R03s. Sometimes they can be the appropriate choice and sometimes the review panel gets it right with respect to criteria and expectations.

  • We just got the scores back and are within the projected payline, funding decision in January! w00t! Thank you America!

  • Mr. Gunn says:

    They could just convert R21s into a initial, high-risk grant that pretty much anybody who can write a coherent application can get. That would solve the issue from the other end yet not put new PIs at such a disadvantage applying for a R01 first thing.

  • Mr. Gunn says:

    Those are opinions, and while you're welcome to yours, I didn't see anything that would make me change mine.
    Actually, isn't what I said kinda what you were suggesting in your last paragraph about "administrative review"?

  • DrugMonkey says:

    PhysioProf wrote this particular entry. I don't necessarily agree with him when it comes to R03 and R21 apps..
    yet not put new PIs at such a disadvantage applying for a R01 first thing.
    how so?

  • Mr. Gunn says:

    Ah, sorry, forgot he likes to do the "refer to myself in the third person" shtick.
    There's two points of concern: The first is the amount of preliminary data you need to have in a R01. The second is the bias (or completely reasonable risk aversion if you prefer) towards established & esteemed labs and away from new PIs, unless they have support from an esteemed lab. Blanket funding/administrative review would help balance the bias scales, even if nothing can be done about the preliminary data situation.

  • DrugMonkey says:

    Preliminary data is a culture, not a requirement and if study sections were responding properly to instructions about n00bs this would not be a problem. I really disagree with doing BandAid end runs around problems because you refuse to face them straight up.
    It is unclear to me what starter grants do to fix the situation. The n00b Investigator who has completed an R21 is still not competitive with one who has completed one R01. The money is less and the duration is shorter. It is very hard to show productivity and maybe even impossible with some models. So if you shift to a defacto requirement for a starter grant you may pick up a broader number of investigators, but you are going to reduce the diversity.
    It is far better to have a system that fits the mechanism to the project instead of to the investigator stage.

  • Mr. Gunn says:

    OK, so I'll defer to your experience on the preliminary data issue, but what about institutional bias/completely reasonable risk aversion?
    Is it important to try to incentivize good ideas, even if they don't come from a well known lab?

Leave a Reply