Which institutions may be hardest hit by the proposed NIH funding cap?

Which institutions may be hardest hit by the proposed NIH funding cap?

At the recent NIH Council of Councils (viewable in this webcast), Michael Lauer presented the proposed cap on the NIH grants using the Grant Support Index (GSI) (see our previous post on the cap) and the presentation was followed by an active discussion. The talk begins at 1 hour 23, and the subsequent discussion at 1 hour 57 minutes. The discussion highlights that the cap is constantly being revised, and currently now may affect only 3% of researchers, rather than the 6% suggested earlier. As Lauer points out, 65-70% of NIH-funded investigators are on one R01 or less.

 

Jonathan Epstein, at the University of Pennsylvania, claimed that 70 investigators – he claims their best investigators – at University of Pennsylvania were affected by the proposed cap. Epstein also claimed that two PIs are even moving overseas, discouraged by this proposed move. Epstein also asks, “How many potential Einsteins do we have to lose in favor of this ‘equality for all’ approach that many of us will believe will favor mediocrity in science?”

 

Epstein points out that other things may affect this distribution of funds, such as where in the country investigators are located. So, which institutions are likely to be affected by the cap?

 

Chris Pickett, from Rescuing Biomedical Research, has kindly passed along data gathered from NIH RePORTer. He downloaded all active projects data, sorted by activity code (R01, R21, etc) and inserted point totals for each grant based on Table 1 from the preprint, “Marginal Returns and Levels of Research Grant Support among Scientists Supported by the National Institutes of Health“, by Michael Lauer et al. to associate the number of GSI points each NIH funded PI has.

 

There were 30,624 PIs in the total dataset. 19349, or 63%, have 7 points (an R01-equivalent) or fewer. 1797, or 5.9% of investigators, were above the 21-point cap. These numbers are both close to those provided by NIH, so this data appears to be a reasonable approximation.

 

  • How many PIs per institution are over the cap?

 

There are 250 institutions with one PI or more above the proposed cap, giving a total of 1799 investigators affected. However there is a concentration at the high end of the scale of schools particularly affected. The 16 schools with more than 30 investigators affected are:

 

PIs over GSI cap
UNIVERSITY OF CALIFORNIA, SAN FRANCISCO 57
STANFORD UNIVERSITY 54
UNIVERSITY OF PENNSYLVANIA 48
JOHNS HOPKINS UNIVERSITY 45
BRIGHAM AND WOMEN’S HOSPITAL 41
WASHINGTON UNIVERSITY 40
UNIV OF NORTH CAROLINA CHAPEL HILL 39
UNIVERSITY OF CALIFORNIA LOS ANGELES 38
COLUMBIA UNIVERSITY HEALTH SCIENCES 37
DUKE UNIVERSITY 36
ICAHN SCHOOL OF MEDICINE AT MOUNT SINAI 36
UNIVERSITY OF WASHINGTON 34
UNIVERSITY OF PITTSBURGH AT PITTSBURGH 33
YALE UNIVERSITY 33
UNIVERSITY OF MICHIGAN 32
UNIVERSITY OF CALIFORNIA SAN DIEGO 31

 

  • How many R01 equivalents would need to be forfeited to get all of their PIs under the GSI cap?

 

The again 16 institutions who will have to shed more than 50 R01 equivalents are:

R01 equivalents to shed with GSI cap in place (Points over cap/7)
MASSACHUSETTS GENERAL HOSPITAL 105
UNIVERSITY OF PENNSYLVANIA 104
UNIVERSITY OF CALIFORNIA SAN DIEGO 97
YALE UNIVERSITY 96
UNIVERSITY OF CALIFORNIA, SAN FRANCISCO 91
UNIV OF NORTH CAROLINA CHAPEL HILL 84
WASHINGTON UNIVERSITY 83
STANFORD UNIVERSITY 82
COLUMBIA UNIVERSITY HEALTH SCIENCES 69
UNIVERSITY OF WASHINGTON 65
UNIVERSITY OF PITTSBURGH AT PITTSBURGH 63
JOHNS HOPKINS UNIVERSITY 60
UNIVERSITY OF MICHIGAN 60
UNIVERSITY OF SOUTHERN CALIFORNIA 57
ICAHN SCHOOL OF MEDICINE AT MOUNT SINAI 56
UNIVERSITY OF CALIFORNIA LOS ANGELES 52

Aside: Why are some of the points negative? This analysis may have been complicated by PIs who have multiple appointments. The data were originally processed asking for the GSI points of each PI, who are all on 22+ points. The institutions to which the grants were awarded were then interrogated and the number of points totaled. If points are split across institutions, this may then complicate the analysis.

 

This is a preliminary analysis, but it highlights two important points:

 

  1. Some institutions will be in exceedingly bad shape after this cap, and it is perhaps not surprising that some of these are also the wealthiest institutions. One can take the number of R01 equivalents being lost and divide by 3 to roughly give the number of extra PIs to maintain that number of grants, which is 35 as the most extreme case for Massachusetts General Hospital.
  2. Most PIs don’t have more than one R01 equivalent, let alone 3. Therefore a small percentage of PIs are affected, and are further very highly concentrated in particular institutions. The voice given to this small group should be responded to with this borne in mind, as the responsibility of NIH is to the entire biomedical enterprise. As Epstein himself points out, “Basing NIH policy on correlation is something scientists don’t like to do”, and the same should apply to over-emphasis on extreme data points such as the institutions who have become most reliant on NIH funding. After all, the University of Pennsylvania has the third-highest number of affect PIs, and the 2nd highest loss of R01 equivalents based on the cap, and the bias here must be recognized as part of a scientific discourse on this issue.

 

If these investigators are as excellent as those such as Epstein claim, then they are likely the ones most able to diversify their funding portfolios and seek funding elsewhere, and thus reducing the strain on the system, rather than instead losing promising early and mid-career researchers. It should be remembered that the system is still constantly expanding with more and more early career researchers, still largely being pushed towards the academic career track, and it is the responsibility of the scientific enterprise to ensure sustainability.

 

Stay tuned on the Rescuing Biomedical Research blog for an upcoming post by Chris Pickett on the institutional distribution of K99s and R00s.

6 Comments

  1. Thanks for doing this analysis! I’m curious what mechanism(s) the R01 equivalents are. Is it multiple R01s that drives most of these PIs above three equivalents? Does this change by institution (R01 vs other mechanisms)?

    Reply
    • Good question – let us get back to you on that one!

      Reply
      • I haven’t had a chance to break it down by mechanism, but I can break it down by points. 881 PIs (49%) are over the cap solely because of 7 point grants (including R01, P01, U01, U19). 1,590 PIs (88%) that are over the cap get at least half of their points from 7 point grants. I think it’s safe to say that 7 point grants, likely R01s, are the primary driver.

        Reply
  2. Hi, What is the source of the school data which you use for each University?

    Reply
    • Hi, This is from the NIH RePORTER.

      Reply
  3. “Epstein also claimed that two PIs are even moving overseas, ”

    I guess there is a lot of excellent applicants ready to replace them. 🙂

    Reply

Submit a Comment

Your email address will not be published. Required fields are marked *