The Director's diagnosis, and the machinery arriving to treat it
A bi-weekly newsletter where I share news, opinion, grant opportunities, success tools and tips on personal growth.
The Director's diagnosis, and the machinery arriving to treat it
In May, Jayanta Bhattacharya, PhD, Director of the NIH, published an essay in Imprimis, Hillsdale College's speech digest, adapted from a talk he gave in California on 28 April. It is the fullest account he has given of what he thinks is wrong with American biomedical science and what he means to do about it. Worth thinking about even now, because the mechanisms it describes have started to show up in the NIH Guide.
I will leave aside the false equivalence he draws between how public health officials act during a pandemic and “how science works”; I will note that Dr. Bhattacharya defends the agency he runs. Almost every modern advance in biomedicine traces back to an NIH investment, he writes, and he lists heart disease, cancer, rheumatoid arthritis, cystic fibrosis, and two cures for sickle cell disease. He says he is strongly of the view that NIH is worth saving. Then he argues it has not done its job well for some time, and names three problems.
The replication crisis. Published findings often do not hold up when other people repeat the work. His answer is to fund replication directly and to make it visible: NIH plans to add a replication button to PubMed so you can see whether a paper’s findings have been tested. The Autism Data Science Initiative already funds replication teams. He is blunt that peer review measures editorial approval rather than truth.
Scientific stagnation. Each NIH dollar buys less health improvement than it did decades ago. He backs this with his own bibliometric work: in the 1990s roughly 55 percent of NIH-funded research rested on ideas no more than a year old, and that share fell away after 2000. He pairs it with the age of PIs at their first R01 equivalent NIH award, up from the mid-thirties in 1980 to the mid-forties now, and with a 2019 paper he wrote with Dr. Mikko Packalen finding that willingness to try new ideas declines steadily with career age. He does not contend with scientific revolution, field maturation or mention inflation including BRDPI. NIH has for decades alerted Congress to the loss of purchasing power that is not effectively offset by the majority of the budget increases it has received.
His fix is structural. NIH gets about 100,000 applications a year and funds eight to ten percent. Overall scores, he says, track what he calls the ‘strength of the proposed methods’ score much more closely than the innovation score, so innovative proposals lose. Leaving aside the technical aspects of this claim (is he talking about Factor 1 and Factor 2 and/or prior criterion scores?) if, as he said in the previous section on the replication crisis, “most ideas hatched by scientists do not work or are not true” who can blame peer-reviewers for focusing on feasibility as they try to make the best funding recommendation? In any case. Dr. Bhattacharya has the solution: NIH is retiring what he calls the payline method in favor of theUnified Funding Strategy, under which reviewers and Institutes are free to back work assigned lower scores by peer review and that may fail. (Wait, if most biomedical research results are wrong – see above – isn’t NIH already funding mostly failure?)
Funding concentration. About a third of NIH grant money goes to roughly twenty institutions. He blames indirect costs, which tie investigators to places that already own the electron microscopes and where the newest arrivals are last in the queue for them. His solution is to separate direct from indirect funding so institutions outside the top twenty can compete for facilities support.
That is the blueprint. Themachineryhas started to arrive. On 31 July a page went up on grants.nih.gov for a new activity code, RP1, the Science-Driven Research Project. It covers standard and exploratory research under one code. Exploratory applications need no preliminary data, and anything with a project period of three years or less is automatically flagged as exploratory on submission. Applications where every PD/PI holds ESI status get special consideration, though PD/PIs do not keep ESI status once an RP1 is awarded. NIH is also holding a public webinar on the Unified Funding Strategy on 14 August, promising to explain the rationale and, in its own words, address misconceptions. My first potential misconception might be that this is combining a handful of activity codes (R01, R21 and a few others) into one, for a hard to understand purpose, as shorter, smaller RP1s are apparently handled like R21s and larger, longer ones like R01s.
My comment.
Much of the diagnosis is sound, and none of it is new. The age of first independent award, the pull of methods scores over innovation scores, the concentration of money in a small number of institutions — these have been discussed inside NIH for over twenty years. I have sat in those meetings. Saying it out loud, in public, with mechanisms attached, is more than his predecessors managed. RP1’s automatic exploratory designation is likely a sensible piece of design.
But it is hard to lay it all at the feet of the few who control science.
In my reading of the data, the age at first R01 has a lot to do with the 10-year Early Stage Investigator window, the length of the postdoc and when startup packages run out. I see this every day in my grant consulting: new PIs try to time their first R01 application to around year 8 after their doctoral degree, which is often year 3 of their first PI position, to maximize their preliminary data momentum while still giving themselves time for a couple of cycles. If you want younger recipients of first R grants, make the ESI a sliding scale where earlier submissions are more advantaged than later ones. It's all about how the incentives are built.
The replication crisis is real. My favorite book-length account of it is Rigor Mortis by Richard Harris. Anything that can be done to understand it better and work to improve what goes in the scientific record can only be applauded. It is not yet clear how funding replication studies will affect costs and the rate of progress; I am open to improvements in both, even if at first glance one might think that doing everything two or three times would only increase cost and slow advances. But funding such studies is only one small part of the equation. Remember scientists, like all humans, act within a system, and one of the most important for them is recruitment, promotion and tenure. Will the postdoc whose major contribution was to replicate a study compete effectively for a faculty job? Will the assistant professor with a replication grant get promoted or achieve tenure? Which journals benefit from publishing replication studies? NIH controls nothing beyond the grants.
On a more fundamental level, what bothers me is the distance between the essay’s opening premise and the mechanisms it proposes. It begins by arguing that there should be no scientific authority, that the scientific revolution worked because it took truth does exactly the opposite. The Unified Funding Strategy shifts judgement from the score, a product of the many peer reviewers, to a smaller group of program leaders at NIH and upwards to political appointees and ultimately Institute Directors or even the NIH Director. Paylines are a crude instrument and I have said so, but it is far from clear that the few now being put in control will do better, especially if ideology is rated higher than peer-review based merit. This direct contradiction is not addressed by Dr. Bhattacharya
This is also real. On 4 August, Dr. Bhattacharya wrote to the University of Pittsburgh terminating a health disparities grant in full, on the grounds that NIH is adjusting its portfolio toward the Director’s priorities (. Two weeks earlier a federal judge held that agencies may not terminate awards on priorities identified after the award was made. The same enlarged discretion meant to rescue a bleeding-edge proposal from a mediocre methods score is being used, in the same month, to end funded work.
On indirect costs I think the economics are wrong. Indirect costs are not a prestige subsidy. They reimburse facility expenses an institution has already incurred. Separating them does not automatically flatten the ecosystem. The likely result is that research-intensive institutions absorb the shortfall and everyone else still has no electron microscope. And it doesn’t address the reason for the concentration in the first place, which has a lot to do with critical mass, access to talent and ability to get funded. I myself am curious to see of the changes in review criteria for R grants, implemented just a few years ago, which deemphasize environment will affect the concentration issue (if we get a chance to see it).
A word for the mechanism being displaced. Consensus peer review, and the R01 built on it, are slow, conservative and tilted toward the well-resourced. They also produced all the cures listed approvingly on the essay’s opening pages. Replacing a flawed system that works is a different proposition from replacing one that does not, and the difference is worth stating before rather than after. I will confess that I can’t explain how Dr. John Ioannidis’ 2005 paper, “Why Most Published Research Findings Are False” can be reconciled with the advances in biomedicine we have witnessed over the past 5 decades. And curiosity about causes is entirely absent from anything I have seen come from Dr. Bhattacharya’s mouth or pen.
Bottom Line: The webinar is on 14 August. If the Unified Funding Strategy is going to decide how your next application is funded, that hour is worth your calendar.
Congress puts the OMB grant rule on ice until 11 December
The Senate has passed a continuing resolution funding the government through 11 December, and it carries a provision blocking OMB from implementing its rewrite of the federal grants regulations for as long as the CR runs. The rule had been due to bind on 1 October, the first day of FY2027. The provision also closes a loophole that would have let OMB finalize the rule just before the CR took effect.
It came out of bipartisan negotiation in Senate Appropriations. Chair Susan Collins, who had already called the proposal deeply flawed and asked Vought to withdraw it, said she was pleased the agreement stops it taking effect. Democrats had made the freeze a condition of any short-term funding measure.
Two things this is not. It is not a repeal: the rule survives, and OMB keeps working through the comments. And it is not permanent. The funding deadline and the rule’s fate now share a date, so both get relitigated together in December.
The House takes up the measure in September. The fiscal year ends 30 September.
Bottom line: If you have been putting off working out what the rule means for your awards, you have four more months.
The funding door is closing on visa holders at two career transitions
STAT reported on 7 August that NIH means to restrict the Pathway to Independence award, the K99/R00, to US citizens, permanent residents and noncitizen nationals. Anil Oza’s report is based on an internal NIH presentation, apparently from a working group on the future of K awards, and there is an indication that the K99 may become the KT1 (kareer transition one, I guess). No Guide notice has appeared.
The same change has already been made one career stage earlier, and that part is settled. The NCI F99/K00 predoctoral-to-postdoctoral transition award has been reissued as an NIH parent NOFO, PA-27-037. Under the old NCI announcement there was no citizenship requirement; applicants could hold a valid US visa, with the institution documenting that it covered both phases. Under the new NOFO, candidates must be US citizens, noncitizen nationals or lawful permanent residents by the time of award. Citizenship is not needed to apply, only to receive the money. Other things changed at the same time: three due dates a year instead of one, no institutional application limit, no nomination letter, a K00 phase cut from four years to three, and a K00 salary cap of $80,000.
Together these close the two federal routes by which a scientist on a visa could get from graduate student to independent investigator with NIH support. NIH already restricts career development awards to citizens and permanent residents. The Grants Policy Statement carves out K99/R00 by name as the exception, and the application form’s non-citizen with temporary visa option applies to K99 candidates alone. If the reported change goes ahead, that last carve-out goes with it.
Bottom Line: The F99/K00 change is confirmed and in force. The K99 change is one outlet and an internal document, with no notice published. If you are advising a graduate student or a postdoc on a visa, the first matters today and the second is worth watching the Guide for.
The letter of support is going away
NOT-OD-26-094, issued 27 July, retires the letter of support. When Forms J come in for FY2027, NIH will want letters of collaboration instead, matching the format NSF has used for years.
The difference is between endorsement and commitment. A letter of support has usually been a respected name saying warm things about the project. That is out. Letters meant to endorse an application without an intent to collaborate or provide material support will not be allowed. A letter of collaboration documents an actual working relationship with someone not listed as senior/key personnel or as an other significant contributor. People in those categories already submit biosketches, so letters from them are redundant.
Two specifics worth writing down. Letters should not come from trainees or staff in the PD/PI’s own laboratory, so no graduate students, postdocs or technicians. And once Forms J are in force, NOFO-specific instructions about letters of support are superseded; applications that still include them may be returned without review.
What to do now. Go through the list of people you ask for letters each cycle and sort it into those who will actually do something and those who will say something nice. Only the first group survives. Change the ask as well. The NSF-style letter is short and plain, a sentence or two confirming what the person will contribute, and praise gets in the way. Say that when you write, or you will get a paragraph of warmth you cannot use. And check whether you need a letter at all, because if the collaborator is senior/key personnel you do not.
The Bottom Line: The habit is older than most of our careers, which is why it will be the thing people forget.
