top of page

The Innovation Paradox

Why our biggest science breakthroughs may not be coming from NIH grants


For more from Alexis, please visit her substack at https://alexisogdie.substack.com/


In my last post, I asked: What do taxpayers actually get from NIH-funded research? Today, I want to unpack another theme from the Bhattacharya–Huberman conversation: we’re not innovating enough—at least not with NIH funds.


I’m fortunate to work in a place where incredible things are happening every day. New ideas become therapies, technologies spin out into startups, and discoveries move from bench to bedside. But when I look closely at some of these breakthroughs, many weren’t funded by the NIH. CAR-T therapy is a striking example—it was considered too innovative/risky to get NIH funding early on, yet it revolutionized treatment for childhood leukemia and other blood cancers. This article (and Bhattacharya’s argument) is not a knock on U.S. scientists; the issue lies in how our funding structures reward research.


The Reality of Getting NIH Grants

Let’s take R01s as an example. This is the NIH’s signature grant—37,478 applications were submitted in 2024, and about 7,000 were funded. These awards typically provide $250,000 to $500,000 per year over five years to tackle a significant scientific question. If we include R01-equivalents (similar grants with slightly different labels), we’re talking about roughly 30,000 grants funded annually, accounting for 60–70% of NIH funding.


So what gets funded? Grants are reviewed by three or more peers who evaluate significance (is it important?), innovation (does it shift the paradigm?), and approach (are the methods sound?). Factors like name recognition, institutional reputation, prior productivity, and especially preliminary data also weigh heavily. Often, you need to show you're already partway there before anyone will take a chance on your idea.


Bhattacharya—and many others, including plenty of us who are NIH-funded—have raised a critical concern: if you need to prove you can do exactly what you say before you’re funded, how innovative can the idea truly be? This is the “can you have your cake and eat it too?” problem. We reduce the risk of failure—but we also limit the potential for transformative discovery. In places like Silicon Valley, failure is part of the process. In science, playing it safe might get you funded, but it may not push the envelope.

Thoughts From the Trenches

A few reflections of my own:

a) If failure is a sign of innovation, then I must be wildly innovative. I’ve had plenty of things go sideways—sometimes despite meticulous planning, sometimes due to circumstances I couldn’t control. But that’s part of the scientific process. Failure is how we learn. That said, in academia, failure—especially early on—can derail your career. That’s a real tension.

b) As someone who’s reviewed many NIH grants, I’ve flagged feasibility concerns more than I can count (I may be hiding behind my hands right now…). But I do it because I’ve been there. I’ve tried the thing that didn’t work or required herculean effort to execute. And, I stand by the feasibility question as a valid question. So maybe the question isn’t whether feasibility matters—it’s how we define and weigh it in the context of innovation.


Back to Bhattacharya’s argument: R01s often support incremental steps built on work that’s already well underway. It’s not uncommon for half of the proposed aims to be completed by the time the grant is submitted. That may improve feasibility and reduce risk, but it also steers the system toward safer, more predictable science. This plays out differently in clinical research, where studies often can’t begin until funding is secured, making it harder to front-load the work. And while applied research may be less conceptually innovative, it often delivers more immediate benefits to human health. The challenge is finding a balance between feasibility, impact, and the kind of bold thinking that drives real scientific breakthroughs.


Given the importance of R01s in academic promotion, the incentives are clear: stick to what’s safe and fundable.


The Age Gap and Innovation

Track record is another key component of funding decisions. Established investigators have a distinct edge. The average age for a first R01 is now 44. This is a problem for many reasons but most relevant for this discussion, Bhattacharya’s work shows that older scientists tend to work on older ideas. In the 1980s, the average funded idea was 2–3 years old; in the 2010s, it was 7–8 years old. Interestingly, the one thing that helped? Early-career first authors. When younger scientists took the lead, especially paired with mid-career or even established senior authors, the ideas tended to be newer and more innovative.

(Stay tuned: more to say soon about why we need more robust funding for early-career investigators.)


Groupthink in Science

Another challenge is groupthink. A big-name researcher publishes a high-impact paper, gives talks across the country, and suddenly everyone’s citing it, building on it, and reviewing grants through that lens. Even when the underlying findings later prove difficult to replicate, the field has already moved on. We’ve all seen this play out. This also happens in study section and locally in institutions when mentors downplay the “crazy idea” of their mentees. Disrupting groupthink in science requires deliberate efforts to welcome diverse perspectives, challenge dominant narratives, and reward intellectual risk-taking. That means funding replication and dissenting work, reforming peer review to reduce bias, and creating space for critical voices—especially from outside the usual power structures. True innovation thrives when disagreement is seen not as disruption, but as progress.

Bhattacharya also raised a good point about diversity in terms of geography. Most NIH dollars flow to coastal institutions—often because they have the infrastructure and reputation that funders trust. If we want innovation to thrive, we need to invest in new places, new institutions, and new people—across the country, not just on the coasts. This is a little hard to say as I’m at one of those coastal places… but being from the Midwest, having scientists funded by the NIH locally could also impact how people in the “middle states” view research.


Let’s Talk Incentives

All of this reflects a deeper structural issue: the incentive system in academic science. Promotion still hinges on securing R01s—which tend to favor safe, incremental work—and publishing in high-impact journals, which sometimes prioritize trendy topics over transformative ones. Metrics like the H-index reward longevity more than innovation, giving an edge to senior investigators with years of accumulated citations. And the prevailing “rockstar” model of science still values individual achievement over collaboration, pushing researchers to advance their own agendas rather than engage in true team science. (More on that soon.)


Where Do We Go From Here?

We need to create:

  • More space to take risks

  • More support for early-career investigators

  • More opportunities for institutions outside the traditional power centers

  • More tolerance for failure—especially when it comes from trying something bold


Innovation doesn’t happen when we reward only what’s predictable. It happens when we give people the freedom—and the funding—to take big swings.


So what does that look like in practice? It currently means diversifying funding sources—disease-specific nonprofits and private sector entities tend to have more risk tolerance. It also means rethinking how we evaluate feasibility, how we score innovation, and how we learn from failure. I have always encouraged my mentees to pay particular attention to that section in the grant “potential pitfalls and alternative strategies” because it tells the reviewer what you get out of the grant in the case of failure. Maybe this is a potential area for augmentation in grants. Personally, I’ve learned far more from the projects that failed than the ones that went smoothly (though I’ll admit, smooth ones are more fun…).

Up next: the “rockstar” model in academia, why funding early-career investigators matters more than ever, and the critical role of replication and reproducibility in science.


References

Age at first PI grant:

Bhattacharya paper on the Age of Ideas:


My disclosures: I am an academic rheumatologist, epidemiologist, and mom. My research is funded by the NIH, private foundations, pharmaceutical companies, and philanthropy. I consult for and work with pharmaceutical companies in my research. I am co-founder of a non-profit organization and founder of Research Pathfinder, LLC. My thoughts are my own and not reflective of my employer.

 
 
 

Comments

bottom of page