Restructuring the NIH and its grant programs to ensure stable careers in science

It is an amazing time to do science, but an incredibly difficult time to be a scientist.

There is so much cool stuff going on. Everywhere I go – my lab, seminar visits, meetings, Twitter – there are biologists young and old are bursting with ideas, eager to take advantage of powerful new ways to observe, manipulate and understand the natural world.

But as palpable as the creative energy is, it is accompanied by an equally palpable sense of dread. We are in one of the worst periods of scientific funding I – and my more senior colleagues – can remember. People aren’t just worried about whether their next grant will get funded, they’re worried about whether a career in academic or public science is even viable (see Kate Clancy’s excellent post on the subject).

There seems to be a broad consensus among the leaders of our community, such as they are, that the solution is for Congress to give us (them) more money. I get emails or calls every few days urging me to contact my senators and representatives to urge them to increase the NIH budget. While I am, in the abstract, in favor of more money for research, if I were in Congress and Francis Collins came to me asking for more money, I’d say “I’m happy to bolster our support for scientific research, but we’re not giving you another single dime until you get your s**t together and stop using the taxpayer’s money to patch over bad decisions and bad policies.”

There are so many things wrong with the NIH today, I could write a book. It’s become an immense, bloated bureaucracy that’s lost sight of its central missions. If it were up to me I’d break it up – turning NIH intramural research into a stand alone entity and creating a separate Institute for Basic Biomedical Research charged allocating funds currently under NIH control to support outstanding and innovative research and to ensure that stable training and career paths exists for American scientists.

This latter issue is the one I want to focus on here. Despite all the challenges of the moment, a lot of outstanding work is getting funded. The problem is, outstanding science needs outstanding scientists. And a lot of outstanding scientists, especially young ones, are leaving academia, unwilling to spend their lives chasing – and in all likelihood not getting – grants.

If I were put in charge of this new institute (or the existing one for that matter) I would devote a large fraction of my budget (I think $10b a year would be a good start) to a “career” award program (not to be confused with the NSF’s CAREER awards).

I would put ~$1b into a pool for young investigator awards. These would be somewhat like current K-99s, in that they would primarily awarded to senior postdocs. These would provide modest startup funds and research support of ~$150k/year for six years – allowing researchers to establish their independent research programs without having to worry about grants. There would be a lot of these – on the order of 1,000 per year. These grants – which would be allocated on the basis of a “people not projects” review, and in all likelihood universities would compete to recruit soon to be independent scientists with these awards.

Recipients of these awards would be evaluated after five years in much the same way people go through tenure reviews today. The purpose of the review would be to assess the researchers contributions to the field and potential for further success. Some would fail to advance, others would be placed in to one of five tiers, representing annual support of between $100,000 (tier 1) and $500,000 (tier 5) – most would be in tier 2 or 3. Every three years research in the career system would be evaluated, with the result of an assessment of their work leading to then either staying in the same tier or moving up or down at most one tier. The total number of people in each tier would be fixed.

I will confess this idea was heavily influenced by the way European soccer leagues operate. At the end of every year, the top teams in each league are promoted to the next higher league, the bottom teams are relegated to a lower division. The system provides a clear opportunity for advancement, but buffers declines – people would only lose their funding after a prolonged period of poor performance, rather than precipitously as happens in the current system if grants do not get renewed.

I estimate that this would cost around $7b/year including overhead. The remaining $3b would support a pool of ~4,500 postdoctoral fellowships and ~12,000 graduate fellowships for trainees to work in career scientists labs. These numbers were meant to provide a pool of 1,500 rising faculty candidates and 2,000 new Ph.D.’s every year, my estimate of what it would take to continually replentish the system.

The $2b left would support a robust equipment grant program for career scientists, including core facilities at institutions with appropriate numbers of career researchers. If the powers that be decide we need more (or fewer) scientists, you scale the whole system by adding or subtracting slots in proportion to available funds.

The $10b was specifically meant not to take the entire NIH extramural budget, but to leave room to fund specific projects, especially high-risk/high reward ones from either career or other labs.

The main goals here are to separate the two crucial function of our granting systems: 1) to fund cutting edge science, and 2) to support a robust scientific infrastructure by providing stable careers to our successful scientists. As I’ve said before, (1) requires (2), but one of the most significant pathologies of our current system is that we mix the two together. In order to support their ongoing research operations, scientists are compelled to dream up “innovative” new projects that can sell in study sections, but often don’t make sense in the real world, while at the same time avoiding truly innovative projects for fear they will be penalized. If labs have a separate mechanism to ensure their financial stability, they will both have more bandwidth to dream up and implement new projects, and the freedom to aim for the stars without worrying they will end up on the street.

I’m sure there are a lot of things I haven’t thought about here, and countless details that need to be dealt with. And I’m equally sure that a lot of people will hate this proposal. But I wanted to put this on the table and open it up for discussion, because the one thing we can not do is nothing. We are dangerously close to losing a generation – or many generations – of scientists. Let’s figure out how not to let this happen.


Addenda: Commenter Jonathan below misunderstood the number of people who would be supported under this system. This was not meant to be an exclusive program. I based my numbers of ~1,000 PIs enter the system per year, with a steady state number probably around 15-20,000. This was a back of the envelope calculation taken from the current size of the NIH grantee and trainee pools. The idea was to stably support a pool of scientists roughly the same size as the current NIH grantee pool, with the PIs trading a more stable funding situation in exchange for lower average levels of support.

This entry was posted in science. Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. Posted January 28, 2013 at 4:48 am | Permalink

    I’m not in bioscience, I’m a psychologist, but as an early career researcher who’s wasted a lot of time hunting for that one magical grant I wholeheartedly agree with the concept of investing in early career stability. It takes time to get a lab established, there are lags at every stage from idea to publication and these are just the way it is. Stable performance based funding without the risk of *abrupt* funding loss is exactly what’s required to move from post-doc to independent researcher.

  2. Jonathan
    Posted January 28, 2013 at 6:08 am | Permalink

    It’s early here so my maths might not be 100% spot on, but you want to pay for ~1000 PIs, 5000 postdocs, 12000 grad students – what are you going to do with the other tens of thousands of PIs and the ~95000 other postdocs that don’t fit into the Eisendrome? Melt them down for glue? Soylent Green?

    At least you’ll cure the problem of for-profit journals, most of them will collapse as there won’t be enough volume of work to keep them open.

  3. Joanna
    Posted January 28, 2013 at 6:38 am | Permalink

    I like the general idea, but where does clinical work fit into this scheme? Surely clinical trials should be assessed by the promise of the project. The research team simply needs to be competent and not too plagued by conflict of interest, and clinical trials should not be changing direction in midstream. I would actually split the NIH up along clinical vs. basic research lines, because the two are really so different, to allow each to find systems to optimize their rather different methods and aims. At that point I don’t see why NIH-basic-research should be treated differently from NSF, so why not merge the two and apply something like your scheme to the new entity.

  4. Jonathan
    Posted January 28, 2013 at 7:52 am | Permalink

    A follow up thought occurs. If this work would remain extramural, you’ll need to add another 50% for indirect costs to the institutions, so you’ve now spent half the NIH budget on a tiny pool of handpicked researchers. Explain to me how this is better than paying it to a much larger pool of handpicked researchers?

  5. Posted January 28, 2013 at 8:08 am | Permalink

    Jonathan – that’s 1000 PIs entering the system per year, steady state would be something in the tens of thousands. The idea was to offer a lower average level of support to investigators in exchange for more stable funding and thereby support MORE scientists, not fewer.

  6. Arash Komeili
    Posted January 28, 2013 at 8:53 am | Permalink

    I like it but I think three year reviews are too short. This, from what I hear, is one of the issues a lot of people have with NSF. By the time the funds arrive, you hire a postdoc or grad student and get them acclimated to the system a significant amount of time has passed (lets say a few months). Depending on the frequency of review panels, you may then have to write the renewal a year and a half into the grant. How will productivity be assessed then? I would favor a 4 year review period. Or fundamentally change the review system. For instance have 5 review cycles a year with faster turn around times.

  7. Posted January 28, 2013 at 8:54 am | Permalink

    Changing the review system is another important thing that needs to be done. Topic for a later day.

  8. Angela DePace
    Posted January 28, 2013 at 9:56 am | Permalink

    I’m all in favor of something that provides long term stability to smaller groups. But how do we make a large scale restructuring of the NIH happen? Seems like its not even on the table at the moment. The types of changes I’ve heard about (funding more early career awards, changing the format of the biosketch, etc.) are tiny band-aids on a much deeper problem.

  9. DrugMonkey
    Posted January 28, 2013 at 9:59 am | Permalink

    One size does not fit all types of science. What a medicinal chemist does with $150K versus a rat lab versus human subjects research in substance abuse is dramatic. Your plan runs the risk of narrowing everything down to bench jockey science because it is cheapest. Therefore productivity will be highest, therefore reviews will go well, therefore…

    And because this process never stops, the highest funded tier will not be as free as you think. The entire lab will be conservatised by the need to show productivity to keep up with the Joneses. It brings the competition over funds to a single unitary focus….beat the Jones lab to keep from being relegated.

    Do you anticipate keeping the IC structure?

  10. Posted January 28, 2013 at 10:12 am | Permalink

    Most importantly, each supported lab should have a colorful uniform that every lab member wears, and there should be continuously updated rankings of lab productivity based on a time-weighted integral of publications x journal impact factor.

  11. Posted January 28, 2013 at 10:13 am | Permalink

    Ready Scientist One

  12. Posted January 28, 2013 at 10:27 am | Permalink

    This is an interesting idea, because I think it starts a question – What is it that we want from our funding agencies? If we can define that, we can begin to think about how can we evolve the system to get there. This spares the typical rants about how the systems is broken, and rather gets to the point where we can debate solutions.

    Let’s not just throw out the idea that more money is a bad thing. The primary cost to me, as a lab director, is my personnel. Since 1998 the money set aside by NIH to fund R01s has approximately doubled (from $1B to $2B – data from Sally Rockey blog), but so have salaries for graduate students and post-docs. In essence, as we’ve trained more scientists funding in constant dollars has flat lined.

    So if we want funding agencies to ensure the next generation of scientists, we need more money. And, if so, we might want to think about an NIH-style institute whose goal is to do only that, focus on training. They can focus on best practices of training, etc., and therefore decouple costs for trainees from investigators. I’m thinking funds from this body would be awarded to institutes, rather than individuals, but we could have a more complex model. I think this does two things. One it stops punishing students who may have had the misfortune to join a lab whose funding is running out. Second, it eliminates the need of the investigator to ask for money to fund their personnel, and allows them to focus on the science.

    One more thought, I’m also concerned using funding as a fulcrum to drive innovation. I agree we want people to be thinking creatively, but if you look at history, necessity has done a better job spurring innovation than dangling a money prize. For example, Sanger sequencing has revolutionized our ability to connect phenotype to genotype. The technique has since evolved, including into many of the whole-genome sequencing platforms available today. That advance came from a need, where people thought creatively about solutions, not because they needed an innovative angle for their grant. I’m curious whether anyone has evidence that requiring innovation actually produces better science.

  13. lazybratsche
    Posted January 28, 2013 at 11:46 am | Permalink

    Who does the actual bench work in this system? If I’m understanding you correctly, each PI would have just one grad student or postdoc. I’m absolutely in favor of having sustainable training and career schemes. But this scheme would dramatically cut the number of people doing the actual work. Would they be replaced by professional technicians?

    Actually, that sounds like a pretty great idea to me. There are a lot of mediocre grad students and postdocs who might flourish as super-techs, if it was a reasonable career path.

  14. Jim Woodgett
    Posted January 28, 2013 at 11:49 am | Permalink

    Thanks for putting some ideas on the table. We need new ideas and not tinkering if we are to provide a compelling ecosystem for researchers as well as not distorting the ingenuity or output of the researchers being supported (this is my biggest fear – that short term thinking is driving short term science, resulting in premature translation and over promise of cures). I like the idea of tiered support with evaluation – as long as there are field adjustments. It also is compatible with the fact that there are non-NIH sources (these might be added to the tier in some manner so there is less likelihood of double-dipping in evaluation). But most importantly, the status quo shouldn’t be on the table as it’ll lead to inexorable decay and an acceleration of the disproportionate loss from younger echelons. This will affect us all as everyone relies on new blood of students and fellows (vampire connotation only partially intended).

  15. Oliver Zill
    Posted January 28, 2013 at 12:51 pm | Permalink

    … “to dream up “innovative” new projects that can sell in study sections, but often don’t make sense in the real world” …

    Speaking as a young geneticist (finished my PhD in 2010), I find this a particularly frustrating attitude/m.o. of many scientists who have “grown up” under an increasingly outdated NIH system. It’s too much about “being clever” and not enough about solving real problems for actual people. Additionally, the obsessive focus on “biochemical mechanisms” to get grants funded and papers accepted for publication feels outdated and stands in the way of important discovery science (pejoratively referred to as “descriptive”) that needs doing right now. In human genetics, for example, it is imperative that we completely characterize the functional effects of as many genetic variants as possible, using both established and new assays. This means there is a TON of fairly obvious but important work to be done ASAP, but much of it would be considered “descriptive” or “stamp collecting” and not “innovative” or “mechanistic.” (Note that I do not think that this particular work should be done via large projects like ENCODE.) If we are truly working for the benefit of humanity, then this work absolutely must be done now. The world has changed and it is time for us to wake up!

  16. dsks
    Posted January 28, 2013 at 2:07 pm | Permalink

    I’m not convinced we necessarily need more PIs (they just need proportionately more PhDs postocs to support, and then we’re back where we started.). It’s also not necessarily the case that more PIs = more and better science anyway. I think tweeking incentives towards improving existing laboratory personnel structures and maximizing productivity in existing labs would be a better place to start. Therefore…

    I’d sooner see recognition at the funding level (and incentives put in place to encourage institutions to recognize it also) that modern productive research labs MUST operate on some form of pyramid structure; competent project directors overseeing a diverse and skilled workforce. Instead of throwing money at graduate programs and new investigator programs, throw money at beefing up and stabilizing the mid career staff scientist positions. Tax payer’s money goes down the shitter every time good bench scientist is forced off the end of the career conveyor belt, not due to a lack of competence, but simply due to not being deemed worthy of new investigator status.

    I’d say divert money currently going towards research assitantships and training grants towards mid/senior postdocs and research assistant profs in an attempt to encourage PIs to retain their best staff (with the proviso that “good staff” is at least in part defined by productivity, which will have to be defended in the budget justification; i.e. no money for 7 yr postdocs with 1 first author and a couple of middle author manuscripts). I don’t think major pay rises are called for for these positions incidentally*, I think the current sallary scale for postdocs is reasonable. The problem is that, with benefits, that compensation is tearing big chunks out of current PI funding and hampering productivity by reducing expenditure on equipment/supplies and whathaveyou. This reality, of course, further makes any solution involving increasing PI numbers seem a bit crazy; we can barely support the ones we have now.

    * I’m also for allowing for a bit of flexibility in salary to allow staff to be rewarded for productivity. Many will wail that this simply incentivises fraud, as if the current do-or-die system doesn’t do that already. Cerainly the incentive to fake data for an extra thousand bucks a year is a lot less than faking it to save your entire career.

  17. Dave
    Posted January 28, 2013 at 5:06 pm | Permalink

    I don’t know. The K99 program hasn’t exactly worked in the way that everyone had planned. I think partly because the NIH has gutted it to keep R01 paylines relatively stable in the past few years, but mainly because those who typically get these grants are already in a position to be competitive for TT jobs, R01s etc. So I’m not sure what these grants achieve at this very moment, but I do agree that career awards are quite valuable overall in the various flavors that they come in.

    The problems are much bigger than this though and you hit the nail on the head when you discussed whether an increase in the NIH budget would be a good thing. Right now it is probably the worst thing that could happen in terms of the future of NIH funded research. If it happened the NIH would just continue the status quo and continue to rely on year-on-year increases to feed the increasing number of mouths. The NIH MUST stop pretending that there is not a major problem and they must stop talking about how great it is that R01 success rates are the same as last year or the year before etc etc. They miss the point over and over again.

    I’m just not convinced that the current NIH is prepared to deal with years of flat funding and I’m worried that they will not make bold decisions when they really need to.

  18. Dave
    Posted January 28, 2013 at 5:10 pm | Permalink

    As far as ideas, at this point I am beginning to favor a baseline funding rate for all/most investigators with “add-on” grants available. Obviously this baseline funding would not be huge, but it would keep the PI in a job and give him a few pennies to keep things ticking over. Luxuries like NCATS, consortia (including ENCODE) and other pet programs have to go.

  19. Ian Holmes
    Posted January 28, 2013 at 6:08 pm | Permalink

    To really make it like soccer leagues you’d need constant intra-league competition. Like an online seminar (e.g. phyloseminar), but two speakers back-to-back, and viewers vote on whose research is coolest. Research A/B testing.

    Seriously though, this would be awesome. I may have spotted a tiny flaw, in that you do not (currently) run the NIH. How do we change that?

  20. Jacob C
    Posted January 28, 2013 at 11:41 pm | Permalink

    (apologies if this posts twice; I forgot to disconnect from VPN when first posting)

    Funny coincidence that this proposal+discussion are taking place now. Just yesterday I was having a very similar conversation with my better half.

    Consider me a casualty of the unstable funding environment.When I went on the job market a few years ago, I was strongly encouraged to take the academic track, but didn’t like what I saw when extrapolating the academic funding fubar 5-10 years into the future. In the end, I was fortunate enough to be offered a position as a group leader doing basic research at an innovative biotech that publishes like crazy. And so, after much deliberation and mental gnashing of teeth, I left academia.

    I’m an evangelical advocate for the role of academic science, and I often think about going back “into the fold” of academia (mainly for philosophical reasons). But when I talk to friends who are 2-3 years into an assistant professorship, I am always warned away by their crushing stress about funding.

  21. Spiny Norman
    Posted January 29, 2013 at 12:28 am | Permalink

    Dave is going to start by gutting NCATS and ENCODE, and other pet programs. As Dave marches to battle I’ll be right with… er, right behind him, all the way!

  22. Jonathan
    Posted January 29, 2013 at 6:52 am | Permalink

    “One more thought, I’m also concerned using funding as a fulcrum to drive innovation. I agree we want people to be thinking creatively, but if you look at history, necessity has done a better job spurring innovation than dangling a money prize. For example, Sanger sequencing has revolutionized our ability to connect phenotype to genotype. The technique has since evolved, including into many of the whole-genome sequencing platforms available today. That advance came from a need, where people thought creatively about solutions, not because they needed an innovative angle for their grant. I’m curious whether anyone has evidence that requiring innovation actually produces better science.”

    Yes, it came from a need. It also came from the same NIH and the same large science programs being decried here.

  23. Jonathan
    Posted January 29, 2013 at 7:04 am | Permalink

    What stops the three-year review process from just turning into a slightly different variant of what we have now with R01 reviews? You’d have to have the same sort of federal government bureaucracy managing the process (because you’re spending public money).

  24. Posted January 29, 2013 at 7:07 am | Permalink

    Mike, you know I’m a big fan of your zeal on all issues relating to the direct and collateral damage of the Tenure Games. I’m part of the aforementioned lost generation of 30something scientists (post-postdocs) that is facing the lousy choice between spinning my wheels in a second postdoc — in my case after managing my own group for 5 years with a $1M budget as an independent fellow — or exiting Academia.

    However, what I thought was missing from your otherwise reasonable proposal is the notion that NIH could help to mobilize independent young and established scientists to do basic and applied biomedical research outside of the university. I would like more free agency in biomedical research, and the NIH is not bound to fund academic PIs exclusively.

    If we’ve actually entered a period of serious experimentation with the way we do science and all options are on the table, then we should be brainstorming how to marshal the talents of life scientists in hybrid laboratory settings that are neither 100% Academia nor 100% Industry.

  25. Dave
    Posted January 29, 2013 at 8:59 am | Permalink

    Yeh thanks Spiny. I always knew you would have my “back”. Coward 🙂

  26. EvoStevo
    Posted January 29, 2013 at 11:22 am | Permalink

    Part of the problem in Science is that the technical track and the managerial track are the same track. Particularly in Bioscience, the Masters degree is under utilized by the discipline as a career trajectory. Engineering and several other fields have masters/certification structures that provide a workforce that is separate from the managerial/conceptual leadership.

  27. EvoStevo
    Posted January 29, 2013 at 11:32 am | Permalink

    Also, the fact that you cannot get a job with a B.S. in biological sciences says volumes about the way our profession handles undergraduate preparation.

  28. Posted January 29, 2013 at 9:03 pm | Permalink

    Michael, good discussion. Unfortunately, for scientists, solving our problems is difficult because we generally don’t communicate very well. Fortunately, #Scio13 exists.

    And we’re poor managers of money (in school we were taught science, not business). So I agree with you, why would anyone give us more money until we learn to do a better job of managing it.

    Also, when you say “bad policies”, I agree. But let’s also add “policies that don’t exist.”

  29. allison stelling
    Posted January 30, 2013 at 4:06 am | Permalink

    i was having a facebook conversation last week about “prestigious” USA universities fighting over the funding with a college buddy of mine who’s a neuroscience PhD student in Boston about this problem (see below for my comment to her). i do think this is part of the current problem with how the NIH allocates its rewards. i for one am cheering you on in this re-structuring, but i do think you’ve got your work cut out for you.

    ” i’m really more worried about “curing cancer”— even just to diagnose it more accurately will require major level, big lab research. you’d want to evenly distribute all grant money through the entire country, so experiments can be replicated and verified in parallel. right now the west coast schools are leading for NIH cash (which is how i personally rank schools atm), with the boston/new york area not too far behind. all this fighting over who has the biggest grant (or is the most “prestigious”) by the coasts is resulting in the Middle starting to lay off researchers when they should be hiring more, to take care of this small area of the USA as they age and die. there’s a few hotspots— like Ann Arbor– that the feds have been putting money into for a while now that are undergoing rapid expansions right now. if we’re actually going to get anything done about this whole “getting old and sick” problem, we need to stop fighting and start sharing the responsibility. and replicate the hell out of the experiment, so that the stats people stop laughing at me when i tell them my # of n.”

  30. Itsik
    Posted January 30, 2013 at 4:32 am | Permalink

    The proposal has two parts.
    One part adresses junior faculty support. I think that implementing this part would cause universities to reduce startup support by corresponding amounts, de facto pocketing not only the packages they currently offer, but also the indirects. At equilibrium, the young scientists will be at the same situation.

    As for the 2nd part, with all the appreciation I have to European soccer, the impact rankings seem secondary here. Labs shoot for impact anyway, and are financially rewarded for success by better chances for their proposal being funded. The novelty in your proposal is the social security system for funding gap unemployment. Nice idea.

  31. DrugMonkey
    Posted January 30, 2013 at 12:47 pm | Permalink

    right now the west coast schools are leading for NIH cash (which is how i personally rank schools atm), with the boston/new york area not too far behind.

    I am always flabbergasted that people are willing to spout this type of subjective, completely unmoored nonsense when with just a little bit of poking around on RePORTER they can verify the facts for themselves.

    in about 45 sec I generated this list of the top 20 (descending by total NIH funding)

    University of Wisconsin-Madison

  32. allison stelling
    Posted January 30, 2013 at 2:26 pm | Permalink

    DM, i did mention “hotspots”. i just think the funding should be more evenly spread out across rural areas; not just the nice cities. I note most of the schools you mention are in “fairly nice places to live” (according to my mother, at least).

    but, you’re right; i was thinking more about cities with enough of a critical mass of biomed scientists to get solid medical trials done more than anything else. (i just did a biomarker paper for brain tumors, and man, that takes a lot of work and people to get done right.) what i would like to see is replication studies in different labs all over the country (and the world). but, as you so often point out, it’s hard to publish stuff like that in the current environment…..

8 Trackbacks