This image was generated by Google AI (Gemini)

In my last essay I introduced the concept of the forecast graveyard, a growing pile of NIH Notices of Funding Opportunities (NOFOs) that are announced to the research community through grants.gov and then quietly never published. I have now done a deeper analysis of both the forecasts that were formally archived and the forecasts that remain listed as open as of this writing. Of the 35 forecasts that have been archived, 28 remain permanently canceled with no path forward. The picture that emerges is not haphazard. It is a story about two different strategies for eliminating research the current administration does not want to fund, and about a legal obligation to Congress that is being quietly abandoned.

Two Strategies, One Outcome

When people think about NIH cancellations, they typically think about explicit terminations of active, ongoing research grants killed overnight with a form letter. That is one kind of cancellation, and it generated the lawsuits and the headlines.

There is a quieter kind, and it operates through the NOFO forecast system. A NOFO forecast is supposed to announce a forthcoming funding opportunity. In it, NIH tells the research community that a funding opportunity is coming, gives investigators time to prepare, and then provides a projected date when NIH will publish the full NOFO. Instead, NIH forecasts are increasingly functioning as a way to give the appearance of activity while delivering none.

I have compiled and analyzed two lists: 35 forecasts that were formally archived, and 336 that remain open as of today. Of the 35 archived, 28 remain permanently canceled with no path forward, while 7 were eventually reissued in some form. Of those 336 open forecasts, 205 are already past their promised posting date. That is 61% of all open forecasts sitting overdue, with no full NOFO published and no explanation given to the research community waiting for them. The oldest two have been sitting since August 2024, now 581 days in the system. Both are NINDS forecasts for the BRAIN Initiative. One is overdue by nearly ten months; the other by six.

What Got Formally Canceled

The 35 archived forecasts were not all canceled at once, or by the same actors. They arrived in three distinct waves, each reflecting a different phase of the political pressure on NIH.

The first and largest wave came in April 2025, when DOGE swept out 24 forecasts in a matter of days. All carry the following notation in the grants.gov system:

“Archiving forecast pending further agency review.”

When you look at what those 24 have in common, the pattern is not subtle. Eight contained explicit equity, diversity, or health disparities language: funding opportunities for health equity research, diversity in aging, vision health equity, and cancer disparities in people with HIV. Seven involved substance use or alcohol research. Four involved HIV or AIDS. Six were workforce training forecasts. Several hit multiple flagged categories at once. The NIDA “Reaching Equity at the Intersection of HIV and Substance Use” forecasts combined HIV research with equity language and substance use, touching three screened categories simultaneously. These cancellations read like the output of a keyword search, because they almost certainly were.

To be fair, not every April 2025 archived forecast stayed buried. Seven of those 24 were eventually reissued in some form. Three of them, two NIA Alzheimer’s clinical trials and the NHLBI VINYL pediatric lung disease consortium, cleared the approval process and made it all the way to published NOFOs. Four others were reissued as new forecasts and remain in the system, still waiting. The Paul Calabresi Career Development Award in Clinical Oncology from NCI has now been in the forecast system for 19 months across two separate forecast postings, with its expected publication date already 119 days overdue. The NICHD Infrastructure for Population Dynamics Research Program has been in the system for 15 months and is 248 days past its expected posting date. It is worth noting what the three success stories (i.e., the NIA and NHLBI reissues) have in common: they are disease-focused, politically unambiguous, and carry no equity or DEI framing. The forecasts that cleared the process were the ones least likely to attract political scrutiny. That pattern holds even within the small population that got a second chance.

The second wave came in the late summer and fall of 2025, after DOGE had formally departed from NIH. Four more forecasts were archived between August and October 2025 by NIH leadership. These four include the NIH Director’s Early Independence Award, a NICHD grant on bacterial STIs in HIV-affected adolescents, and two NIDCR training programs. Like the DOGE cancellations, these carry HIV and training content, suggesting the political screening criteria were already embedded in agency decision-making before DOGE left.

The third wave is ongoing. Seven forecasts have been archived since February 2026 by current NIH leadership, HHS, and/or OMB. This is where the story becomes more troubling, and the language shift in the grants.gov system is itself revealing. The earlier cancellations said “pending further agency review,” implying a temporary hold and leaving researchers with some hope of eventual publication. The 2026 cancellations carry a blunter notation:

“This NOFO will not be published. The Forecast is being archived.”

No ambiguity, no promise of reconsideration. These include a NIMH forecast for research on crisis response services for suicide prevention, an NIA T32 training program for aging research, an NIA short course for Alzheimer’s disease clinical trial workforce development, an NHLBI forecast for research on clonal hematopoiesis and chronic inflammation, the NIAAA Comprehensive Alcohol Research Centers P60 mechanism, and an NIDCR forecast for dental and craniofacial data analysis.

Only two of these seven contain equity or disparities language. The others were canceled for reasons that are not visible in their content. When forecasts for suicide prevention research and dental research methods are eliminated by the same process, the gate is no longer about flagged terminology. It is about control over what funding opportunities will be allowed.

What Is Being Left to Die Quietly

The more revealing story may be the forecasted opportunities that have not been formally canceled but have not been published either. Many of these appear to be expiring by neglect rather than by explicit decision, and the data support that interpretation.

Consider NIMHD, the National Institute on Minority Health and Health Disparities. All seven of its open forecasts are overdue, by between 122 and 318 days. None have been formally canceled. The NIMHD forecasts involve health disparities research centers, minority institution programs, and the EmbraceHealth Clinical Research Network. These are precisely the kinds of funding opportunities that would attract attention and criticism if explicitly terminated. Letting them sit generates the same practical outcome, opportunities never posted, without creating the paper trail that a formal cancellation would.

NINR, the National Institute of Nursing Research, shows the same pattern. All six of its open forecasts are overdue. All six were posted in May 2025 and cover community-partnered nursing research, health disparities in rural populations, violence against women, and school health environments. They have been sitting for nearly ten months past their expected posting dates.

NIAAA, which posted zero NOFOs in all of 2025, is managing both strategies at once. Four of its forecasts were permanently archived while nine others remain listed as open, seven of them overdue by six to eight months. Researchers in the alcohol science field are sitting with a list of anticipated funding opportunities and no indication of whether any will materialize. The formal cancellations cleared out the most explicitly flagged content. The rest are just sitting.

The distinction between explicit cancellation and passive expiration matters for accountability. A canceled forecast leaves a record. A forecast that simply ages past its promised date never appears in a cancellation tally. For anyone tracking this policy, including Congress, the official record understates the damage by a substantial margin. That understated damage is especially concentrated in one category of funding opportunity that has received almost no public attention.

The Infrastructure Problem

More than half of all open forecasts (192 of 336, or 57%) are for coordinating centers, research networks, multi-site consortia, data hubs, and other large-scale infrastructure programs. These are the programs that build and sustain the platforms on which individual investigators do their work: clinical trial networks, data coordinating centers, SPORE cancer centers, genomic resource centers, and disease-specific research consortia.

Of those 192 center and infrastructure forecasts, 112 are already overdue, with a median overdue time of 166 days. That is notably longer than the 119-day median for other overdue forecasts, suggesting that large programs face a heavier approval burden. The reasons are not hard to understand. A coordinating center involves larger dollar amounts, longer project periods, and more complex programmatic coordination including a special emphasis panel review. Each of those factors adds friction in an approval process that was already slow and is now operating under a political bottleneck.

The consequences for ongoing research are serious. Large centers do not just support investigators applying for new funding. They sustain infrastructure that active researchers depend on right now: specimen banks, data repositories, clinical trial coordination, and training cores. When a data coordinating center’s renewal sits in forecast limbo for 150 days past its promised posting date, the uncertainty ripples outward to every project the center supports. NCI alone has 14 overdue center forecasts, with an average overdue time of 183 days. NINDS has 12. NIA has 10. NHGRI has 10.

Some of these programs have been waiting since mid-2025. The Experimental Therapeutics Clinical Trials Network (ETCTN) Lead Academic Organizations forecast was due in July 2025 and is now 242 days overdue. The Alzheimer’s Clinical Trials Consortium, one of NIA’s flagship clinical research programs, was due in August 2025 and is 241 days overdue. The HeartShare 2.0 heart failure research network from NHLBI was due in October 2025 and is 166 days overdue.

These are not small investigator-initiated grants that can be resubmitted quickly if timing slips. These are programs that require years of development, involve dozens of institutions, and provide the scientific infrastructure that individual investigators rely on for their own work. When they stall, the damage is not contained to the coordinating center. It spreads.

Congressionally Mandated Programs Are Not Exempt

The most legally significant finding in this analysis is the number of congressionally mandated research programs whose forecasts are now overdue with no movement. Congress established these programs through legislation and appropriated funds for them specifically. NIH does not fund them at its own discretion.

The BRAIN Initiative has eight open forecasts that are overdue, ranging from 54 to 301 days past their promised posting dates. The HEAL Initiative, Congress’s response to the opioid crisis, has five overdue forecasts ranging from 63 to 166 days. The Gabriella Miller Kids First Pediatric Research Program has three forecasts overdue by 243 days. This program was named for a child who died of a brain tumor and was established by Congress specifically to fund pediatric cancer and structural birth defect research. The INCLUDE Project for Down syndrome research has three forecasts overdue by 101 days. NIEHS Superfund and HAZMAT worker training programs have four forecasts overdue by 165 to 210 days, including worker health and safety training at DOE nuclear weapons sites.

When NIH fails to publish NOFOs for programs Congress has mandated and funded, it is not simply making a policy choice. Appropriated funds for specific programs that go unspent because the agency will not issue funding opportunities represent a potential violation of the Congressional Budget Act. This is not a hypothetical. It is the structural consequence of a system in which every targeted NOFO now requires approval from NIH/OD, HHS, and OMB before it can be published, and that approval is not arriving.

The Two Populations

The canceled and sitting forecasts reveal two populations of research being curtailed, though the methods used against each have grown more varied over time.

The first population is research with explicit political markers: DEI language, equity framing, HIV research, racial health disparities, and gender-related topics. Most of these were swept out in the first DOGE wave in April 2025, identified by keyword. The pattern was unmistakable and the grants.gov notation left little ambiguity about the intent.

The second population is everything else: basic biomedical science, workforce training, clinical infrastructure, center renewals, and congressionally mandated programs. What is striking is that this population is being curtailed through two different mechanisms simultaneously. Some have been explicitly canceled in the second and third waves, including forecasts for suicide prevention research, Alzheimer’s workforce development, and dental research methods, none of which carry any obvious political marker. The rest are being strangled by process, sitting overdue in the forecast system with no explanation and no path forward. Whether a forecast gets an explicit cancellation notice or simply ages past its promised date, the outcome is the same: research that was announced to the scientific community never materializes.

Every targeted NOFO now requires several rounds of approval both before and during the forecast process from NIH/OD, HHS, and OMB. Two hundred and five forecasts have passed their promised posting dates without a single public explanation from NIH. The forecast graveyard works because it offers no closure, no record of failure, and no accountability.

Researchers monitoring a BRAIN Initiative forecast cannot tell whether their anticipated funding opportunity is stuck in an approval queue, has been silently deprioritized, or is simply never going to arrive. Investigators keep waiting for the official NOFOs to post on grants.gov. Institutions keep planning. And nothing comes.

Of the 336 forecasts currently listed as open, 205 have already missed the date they were promised to the research community. Twenty-four of them are focused on topics that Congress specifically mandated and funded. One hundred and twelve are large centers and infrastructure programs that active research depends on right now. And not one of them has received a formal explanation, defense, or withdrawal.

That is not a funding pipeline. That is a waiting room with no appointments.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

In the service of open science, I have made the data underlying this analysis available for download here: NIH Forecasted NOFOs. This list of NIH forecasts was compiled from grants.gov as of March 30, 2026. Anyone working from the same source of publicly available data should be able to reproduce these findings.

Image credit: @nihvigils.bsky.social

On March 23, 2026, Dr. Jon Lorsch, NIH’s Deputy Director for Extramural Research, published a blog post titled “NIH’s Path to a Simpler Funding Opportunity Landscape.” The post attempts to explain and justify why NIH is issuing dramatically fewer Notices of Funding Opportunities (NOFOs) than at any point in recent history. It is a carefully written document, fluent in the language of scientific stewardship and administrative efficiency. But it is built on assertions without evidence, and it sidesteps the most important facts about what is actually happening to NIH-funded science.

I spent 22 years as a Program Officer at the NIH. I know how NOFOs are created, what they accomplish, and what their absence means for the research community. I have been writing about this policy change on Substack and in The Chronicle of Higher Education, and I have the data to respond.

The Numbers Tell a Different Story Than the Blog

Dr. Lorsch acknowledges that NIH historically issued between 600 and 850 NOFOs annually. He frames the current reduction as a modest, principled streamlining. The data tell a very different story.

Using NIH’s own posting records, which I have compiled across 14 years (2012–2026) from the NIH Guide for Grants and Contracts and grants.gov, here is what the NIH-wide NOFO count actually looks like:

• 2012–2024 annual range: 511 to 1,081 posted NOFOs per year

• Average 2019–2024: approximately 710 per year

• Average 2023–2024: 724 per year (the immediate pre-policy baseline)

• 2025 posted NOFOs: 120, which is an 83% decline from the recent baseline

• 2026 posted NOFOs so far: 17, which is a pace far below even the already-collapsed 2025 level

NIH NOFOs Published Over Time (as of March 24, 2026)

To put 2025 in full context: in 2020, when COVID-19 shuttered laboratories and disrupted normal NIH operations worldwide, NIH still managed to post over 650 NOFOs. The 120 posted in 2025 is the lowest annual total in the entire 14-year record. This is not streamlining. This is a near-total halt.

And the picture worsens when you account for the new “forecasting” system. Under NIH’s new policy, ICOs must first post a forecast (a simple notice of an anticipated future NOFO), before being permitted to publish the actual funding opportunity. Across all ICOs in 2025, 271 opportunities were forecasted but only 120 were ever posted. Nearly seven out of ten forecasted opportunities never made it to the posting phase.

These are not opportunities being carefully considered. I have come to call this the “forecast graveyard.” It is a pipeline that generates false hope for researchers and then quietly buries it. This system introduces a whole new burden on investigators who read forecasted NOFOs that will never come forward. When seven out of ten forecasts never become actual funding opportunities, generating them is largely an exercise in false advertising.

Claim #1: “This Does Not Mean Fewer Funded Applications”

“Although there are fewer funding opportunities, this does not mean fewer funded applications.” — Dr. Jon Lorsch

This is the most consequential claim in the blog, and it is directly contradicted by NIH’s own published data. According to NIH’s own reporting, in FY 2024 NIH awarded 7,720 R01-equivalent awards. In FY 2025, that number fell to 5,885, a decline of more than 20%, despite a nearly flat NIH budget. The link to this data is on NIH’s own website.

Furthermore, according to a recent Association of American Universities report, NIH issued 66% fewer grant awards in FY26 through the end of February compared to the average for the same period in FY21–24. These are not small measurement artifacts. These are catastrophic drops in the actual funding of science, and the opposite of what Dr. Lorsch promises.

It is important to note that the drop in numbers of grant awards did not result from the drop in NOFOs. Rather, as a recent article in Science notes, the switch to a multi-year funding strategy was likely the primary cause. The cause matters less than the consequence. Dr. Lorsch assures the research community that fewer NOFOs would not mean fewer funded applications. The data show that funded applications fell sharply, whatever the reason. NIH is actually funding far fewer awards now, and it appears likely that this trend will continue. Telling the research community that fewer NOFOs will not mean fewer funded applications, while the agency’s own numbers show a 20% drop in R01 awards within a single year, is not a defensible position. It is gaslighting.

Claim #2: Targeted NOFOs Limit Investigator Flexibility

“Their narrow scope can sometimes limit flexibility for investigators to propose novel or unexpected ideas.” — Dr. Jon Lorsch

This is the foundational premise of Dr. Lorsch’s entire argument, and it collapses under the weight of NIH’s own historical data. NIH has tracked, for decades, the proportion of funded research that comes from investigator-initiated versus targeted NOFOs. The NIH Data Book charts show clearly and consistently that broad parent mechanisms — the very investigator-initiated grants Dr. Lorsch says he wants to protect — have always dominated the NIH funding portfolio. Targeted NOFOs have never crowded out investigator-initiated science. This is not a problem that exists.

Dr. Lorsch acknowledges as much, writing that “broad funding opportunities like parent announcements have supported a large share of NIH-funded research for many decades.” If investigator-initiated research has always been the dominant mode, what exactly is the problem being solved?

Here are some representative charts from the NIH Data Book:

In 22 years at the NIH, I do not recall a single instance in which the extramural research community complained that too many NOFOs were forcing them to constrain their scientific ideas. Scientists were always welcomed and encouraged to submit their best ideas through parent mechanisms. Most did. The data confirm it. Dr. Lorsch provides no survey, no listening session summary, no Request for Information, and no documented complaint from the research community to back up his claim that NOFO proliferation was limiting scientific innovation. The assertion is simply stated as fact, repeatedly, but repetition is not the same thing as evidence.

Innovation and discovery are not impeded by the existence of targeted NOFOs. Indeed, the vast majority of Nobel-winning NIH-funded research has been funded through unsolicited, investigator-initiated grant applications during times when more targeted NOFOs were issued at the same time. To date, the NIH has funded 174 Nobel laureates who collectively received or shared 104 Nobel Prizes for their groundbreaking work. All of that occurred during the same decades when NIH was issuing hundreds of targeted NOFOs each year.

Claim #3: Fewer NOFOs Reduce Administrative Burden on NIH Staff

“Streamlining the number of funding opportunities also reduces administrative complexity across NIH, allowing NIH program staff to focus more on scientific stewardship.” — Dr. Jon Lorsch

As someone who spent more than two decades as a NIH Program Officer, I can speak to this directly: the number of NOFOs is not what burdens program staff. Scientific stewardship — talking with investigators about their ideas, nurturing portfolios, identifying scientific gaps — is the job. It is what program officers were hired to do, and it is what they want to do. Having a robust portfolio of targeted NOFOs does not prevent that work. It enables it.

What does create genuine administrative burden and prevent scientific stewardship is something Dr. Lorsch does not mention at all: the ongoing requirement that program staff screen grant portfolios and applications for terms that may potentially be associated with misalignment with the agency’s new priorities’ such as DEI, mRNA research, vaccine hesitancy, and climate research. Program colleagues who remain at NIH describe spending countless hours defending grants in their portfolios, coaching applicants about which words to avoid, and navigating an ever-changing list of politically sensitive topics. This is what diminishes innovation. This is what limits investigator flexibility. This is what keeps program staff from focusing on scientific stewardship.

In 2025, NIH terminated thousands of grants worth billions of dollars. HHS grant terminations impacted more than 74,000 clinical trial participants in the name of eliminating DEI. Even after a federal judge determined that DEI screening of NIH grants was illegal, the practice continues. If Dr. Lorsch genuinely wants to free his program staff to focus on science, he should address that problem.

Claim #4: This Is About “Reducing Fragmentation”

“Moving toward the use of broader funding opportunities rather than highly specialized announcements reduces fragmentation in the funding landscape.” — Dr. Jon Lorsch

Dr. Lorsch never defines “fragmentation,” and this vagueness is telling. NIH has 27 Institutes, Centers, and Offices (ICOs), each with a distinct and congressionally mandated mission. The National Cancer Institute exists to fund cancer research. The National Institute on Alcohol Abuse and Alcoholism exists to fund alcohol research. The Fogarty International Center exists to fund global health research. When these ICOs issue targeted NOFOs that reflect their specific missions, that is not fragmentation. That is the system working as Congress intended.

What the new system actually does is centralize control. Under the previous policy, targeted NOFOs were approved by ICO directors and their affiliated advisory councils composed of scientific experts with deep domain knowledge making peer-reviewed judgments about scientific priorities. Under the new policy, every targeted NOFO must be approved by the NIH Office of the Director, HHS, and OMB — political appointees whose expertise is political, not scientific.

Funneling the research community toward fewer, broader parent mechanisms also makes the grant portfolio far easier to monitor and filter for political purposes. When every application arrives through the same parent announcement, screening for ideologically disfavored topics becomes straightforward. This may not be what Dr. Lorsch intends, but it is the structural consequence of the system he is describing.

The practical result of these new NOFO policies is visible in the data, and it is not evenly distributed.

Which ICOs Have Been Hit Hardest?

The data I compiled reveal that the collapse in NOFO activity is not evenly distributed. Using the average of 2012 – 2024 as the pre-policy baseline, here is what some of the hardest-hit ICOs experienced in 2025:

• NICHD (child and maternal health): posted just one NOFO in all of 2025. Its 14-year average was 41.7 per year, making the decline approximately 98%. Thirteen forecasted NOFOs sit in the graveyard.

• NINDS (neurological disorders): posted 3 NOFOs in 2025 against a 14-year average of 49. That is a 94% decline. What makes the NINDS story particularly striking is the trajectory: the institute had been accelerating its activity for years, reaching what turned out to be its all-time peak of 80 NOFOs in 2024 and then collapsed 94% the very next year. Sixteen forecasted NOFOs remain unposted.

• NCI (cancer): NIH’s highest-volume institute by any measure. Over the full 14-year record it averaged 86 NOFOs per year, and in the five years immediately before the policy it averaged more than 90. In 2025 it posted 4, a 95% decline. Twenty-four NCI forecasts are sitting in the graveyard, representing cancer research programs that were anticipated, announced to the community, and then quietly shelved.

• NIMH (mental health): averaged nearly 60 NOFOs per year over 14 years. It posted 4 in 2025, a 93% decline. Mental health research, an area of enormous and well-documented unmet need, has been reduced to a trickle.

• NIAAA (alcohol research): issued zero NOFOs in 2025 despite having 9 in the forecast system. Its 14-year average was 21 per year, with annual totals ranging from 10 to 44. NIAAA has never had a zero-posting year in the entire 14-year record. Nine forecasted opportunities were announced to the alcohol research community and then never materialized.

• NIEHS (environmental health): posted zero NOFOs in 2025, despite 6 forecasted. Its 14-year average was 14 per year, and like NIAAA it has never previously had a zero-posting year. Calls for targeted research on environmental contributions to disease, including cancer, reproductive harm, and neurological conditions, simply stopped.

The mandate, as I understood it from NIH leadership before I left, was a 50% reduction in the number of NOFOs issued each year. What the 14-year data show is an 83% reduction system-wide, with most individual ICOs experiencing declines of 90% or more.

What Targeted NOFOs Actually Do — And Why Losing Them Matters

Dr. Lorsch’s blog does not spend any time describing what targeted NOFOs actually accomplish. It should. As I explained in my last essay, a NOFO is how NIH tells researchers what the agency needs. It specifies a research problem or a training need, explains why it matters, describes the approach NIH is looking for, and sets aside dedicated funding to solve it. Critically, NOFOs typically focus on problems that are not being addressed by unsolicited investigator-initiated applications. Targeted NOFOs exist precisely to address the gaps that broad parent mechanisms do not fill. When you eliminate targeted NOFOs, you eliminate the funding signal for exactly those areas of science that would otherwise be overlooked.

Moreover, some of the most consequential scientific programs in NIH’s history were built on targeted NOFOs: the Human Genome Project, the BRAIN Initiative, the All of Us Research Program, the Cancer Moonshot, the IDeA program that builds research capacity in historically underfunded states. None of these emerged from investigators independently deciding to pursue them through parent mechanisms. They emerged because NIH identified a strategic priority and issued a targeted funding opportunity to pursue it. To suggest that targeted NOFOs limit innovation is to misunderstand how some of the most innovative federally funded science of the past three decades actually got done.

Many of these programs are also mandated by Congress. The IDeA program, for example, exists because Congress determined that states with historically low NIH funding deserved targeted investment to build research infrastructure. When NIH stops issuing NOFOs for congressionally mandated programs, it is not just making a policy choice. It may be failing to fulfill a legal obligation.

A Personal Example: What a Targeted NOFO Can Accomplish

By 2016, I had recognized that behavioral and social science doctoral training was falling dangerously behind the rapid growth of big data and computational methods. University programs across the country were not preparing graduate students to handle the health data of the future. No investigator was going to solve this problem by submitting a parent mechanism training grant application. It required a strategic, coordinated response from NIH.

I spent years developing a concept, pitching it across institutes, gathering feedback, drafting budget scenarios, and building the case for action. I persuaded OBSSR leadership to commit $2 million per year over five years. I recruited 11 NIH institutes as co-signers. And in 2019, we issued a targeted NOFO (RFA-OD-19-011) to launch the first national data science training program for behavioral and social science graduate students in the country.

The first eight programs were funded in 2020. The RFA was reissued in 2024. Over 100 scholars have now been trained, significantly expanding U.S. behavioral data science capacity in ways that will pay scientific dividends for decades.

That program did not happen because an investigator independently decided to propose it. It happened because a program officer identified a gap, built a coalition, and issued a targeted NOFO. Under the new system Dr. Lorsch is describing, that process would require approval from NIH/OD, HHS, and OMB before a single word of the funding opportunity could be published. Given what we know about how many forecasted NOFOs actually make it to posting (fewer than one in three in 2025), there is a reasonable chance that program would never have existed.

Conclusion

Dr. Lorsch’s blog is well-written, but it is not well-supported. The central claims are each contradicted by available data: that targeted NOFOs limit investigator flexibility, that fewer NOFOs will not mean fewer funded awards, that this is about reducing administrative complexity. The NIH’s own published numbers show a collapse in NOFO activity that dwarfs the mandated 50% reduction, a parallel collapse in R01 awards, and a forecast system in which the majority of anticipated opportunities are never published.

The blog presents none of this data. It makes no reference to the administrative burden created by DEI screening of grants, which is the actual burden consuming program staff time. It offers no analysis to support the claim that investigators felt constrained by targeted NOFOs, because no such analysis exists. It makes no distinction between the policy as stated (50% reduction) and the policy as implemented (83% reduction).

The research community deserves an honest accounting of what is happening. This is not a simplification of the funding landscape. It is a contraction of it, and one that is falling most heavily on the ICOs and research areas that serve the most vulnerable populations, including children’s health, mental health, alcohol and substance use, cancer, environmental health, and neurological disease.

Fewer NOFOs means fewer signals to the research community about what NIH needs. It means fewer protections for understudied areas that cannot compete through open parent mechanisms. It means a transfer of authority away from scientific experts and toward political appointees. And, as the data now confirm, it means fewer funded awards.

That is not a simpler funding opportunity landscape. That is a diminished one.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources

National Institutes of Health, Office of Extramural Research. (2026, February 10). NIH support for early-stage investigators in FYs 2024 and 2025. NIH Extramural Nexus. https://grants.nih.gov/news-events/nih-extramural-nexus-news/2026/02/nih-support-for-early-stage-investigators-in-fys-2024-and-2025

Agarwal, K. (2026, March 20). Data show dramatic slowdown in NIH grantmaking. Association of American Universities. https://www.aau.edu/newsroom/leading-research-universities-report/data-show-dramatic-slowdown-nih-grantmaking

Kaiser, J. (2025, February 20). NIH research grant funding rates plummeted in 2025. Science. https://www.science.org/content/article/nih-research-grant-success-rates-plummeted-2025

Patel, V. R., Liu, M., & Jena, A. B. (2026). Clinical trials affected by research grant terminations at the National Institutes of Health. JAMA Internal Medicine, 186(1), 126–128. https://doi.org/10.1001/jamainternmed.2025.6088

National Institutes of Health. (n.d.-a). Research project success rates [Report 25]. NIH Data Book. https://report.nih.gov/nihdatabook/report/25

National Institutes of Health. (n.d.-b). Research project grant (RPG) awards by NIH Institute/Center and mechanism [Report 26]. NIH Data Book. https://report.nih.gov/nihdatabook/report/26

Wilcox, C., Richter, H., Voosen, P., & Kaiser, J. (2025, December 12). Do you wanna build an ice dome? We could live in there on Mars. ScienceAdviser [Newsletter]. Science. https://www.science.org/content/article/scienceadviser-do-you-wanna-build-ice-dome-we-could-live-there-mars#nih-grant

National Institutes of Health. (2025, December 12). Staff guidance: Reviewing grants for priority alignment [Internal document obtained by Science]. https://www.science.org/pb-assets/PDF/News%20PDFs/final_staff_guidance-1765918233.pdf

Ginexi, E. (2026, March 17). Why is the NIH abandoning science? This is not a temporary slowdown. The Chronicle of Higher Education. https://www.chronicle.com/article/why-is-the-nih-abandoning-science

NOFO data was compiled by the author from NIH Guide for Grants and Contracts historical archives and grants.gov records. Analysis covers all funding announcements published from 2012 through March 24, 2026.

In 2024, NIH published 756 funding announcements.

In 2025, it published 120.

In 2026, as of March 15, it has published 14.

I spent 22 years as a program official at NIH writing these announcements. I know what they accomplish and what happens when they disappear. This essay is about what the data reveals and what it means for every disease, every research area, and every population that depends on NIH-funded research.

Figure 1. NIH NOFOs Published Over Time

What Research Funding Announcements Actually Do

A Notice of Funding Opportunity (NOFO) is how NIH tells researchers what the agency needs. It specifies a research problem, explains why it matters, describes the approach NIH is looking for, and sets aside dedicated funding to solve it.

NOFOs exist because not all research needs are obvious to individual investigators. When a new pathogen emerges. When clinical trials reveal an unexpected side effect that needs investigation. When one population experiences a disease at higher rates than others but nobody knows why. When a promising scientific approach exists but no one is applying it to a specific problem. These are moments when waiting for unsolicited grant applications is not enough.

Writing a NOFO was one of my primary responsibilities as a program official. When my institute identified a gap, I would work with scientific experts to define the problem precisely, determine what kind of research was needed, and draft an announcement that would attract the right investigators. The announcement would be reviewed by our advisory council, posted publicly, and researchers across the country would know: NIH has identified this as a priority and has set aside funding to address it.

This is scientific stewardship. It is not top-down control of what researchers can study. Investigators can always submit unsolicited proposals on any topic within an institute’s mission. But NOFOs allow program staff to actively direct resources toward problems that need attention rather than passively waiting to see what applications arrive.

And then it stopped.

The Collapse

I downloaded all available NOFO data from NIH’s historical records and grants.gov and analyzed every funding announcement published from 2012 through March 15, 2026. What I found was worse than anything reported in the media.

From 2012 through 2023, NIH published an average of just over 700 NOFOs per year. There was variation (a peak of 1,110 in 2017, a low of 535 in 2012), but the system was stable. Institutes identified research needs and issued announcements to address them. In 2024, that number was 756, still within the normal range.

Then the collapse began. By 2025, only 120 announcements were posted, an 84% decline. As of mid-March 2026, only 14 have been published. If the current pace continues, this year will see a 98% reduction from historical norms.

The decline is not limited to a few institutes. It is systemic. The National Cancer Institute, which historically published more NOFOs than any other institute, has gone nearly silent. The National Institute of Allergy and Infectious Diseases, responsible for pandemic preparedness and emerging disease response, has published almost nothing. Institutes focused on mental health, aging, drug abuse, environmental health, and rare diseases have all but stopped issuing targeted funding announcements.

This is not a temporary slowdown. It is a structural collapse.

Figure 2. NIH NOFOs Published by Each Institute/Center Over Time

The Policy That Caused This

The collapse began with a policy change. In recent guidance, NIH leadership announced an “overall reduction in number of NIH NOFOs.” The stated goal was to streamline funding opportunities and reduce redundancy.

At the same time, the approval process for NOFOs was fundamentally restructured. Previously, funding announcements were primarily reviewed and approved within individual institutes by scientific program staff and external advisory councils composed of researchers and public representatives. The process typically took weeks to a few months.

Under the new system, every NOFO must be approved by political appointees in the NIH director’s office and the Department of Health and Human Services before it can be posted. The approval process now takes a minimum of six months, and many announcements seem to be remaining in review indefinitely.

More recently, NOFOs must also be approved by the Office of Management and Budget. This adds another layer of political review to what was previously a scientific decision-making process. OMB now has effective veto power over which research priorities NIH can pursue.

Additionally, NIH stopped publishing funding announcements directly through its traditional NIH Guide for Grants and Contracts. Instead, all NOFOs must now be entered into grants.gov as “forecasts” and wait for approval before becoming active funding opportunities.

The effect of these changes is visible in the data. Institutes are still identifying research needs. Program staff are still writing announcements. But the announcements are not being approved. They sit in forecast limbo, written but never posted, planned but never executed.

The Forecast Graveyard

This is where the data becomes particularly damning.

In 2025, NIH institutes entered 391 total funding announcements into grants.gov (they all start as forecasts). Only 120 of them were actually posted and opened for applications. That means 271 announcements (69% of what was planned) were written, reviewed internally, entered into the system, and then blocked at the final approval stage (they still appear as forecasted).

In 2026, so far institutes have forecasted 75 announcements, and only 14 have been posted. Sixty-one remain in limbo (81% of what was planned).

These are not ideas that were considered and rejected for scientific reasons. These are fully developed funding announcements that passed internal scientific review, were deemed important enough to allocate budget toward, and were ready to go. They are sitting in a bureaucratic queue waiting for political approval that, in most cases, never comes.

The forecast graveyard proves this is not about scientific prioritization or budget constraints. If it were, the announcements would not have been written and forecasted in the first place. This is about centralized control. About requiring that every research priority, every identified gap, every targeted funding decision be approved by political appointees rather than scientific program staff. And the result is paralysis.

Figure 3. NIH NOFOs in Grants.gov Under the New “Forecast” Policy

What’s Being Lost

When NIH stops issuing targeted funding announcements, specific kinds of research become much harder to sustain.

Rare disease research suffers because individual investigators are unlikely to propose studies on conditions affecting small populations unless NIH signals it is a priority and has dedicated funding. Research on health disparities struggles for the same reason. Studies requiring particular methodologies, specific patient populations, or coordination across multiple sites all depend on NOFOs that describe exactly what NIH is looking for and commit resources to support it.

Emerging threats become harder to address quickly. When COVID-19 emerged, NIH issued emergency funding announcements within weeks. Those NOFOs allowed the agency to mobilize researchers rapidly toward specific problems: vaccine development, therapeutic testing, long-term effects, vulnerable populations. Without the ability to issue targeted calls, response to future health emergencies will be slower and less coordinated.

Innovation in underfunded areas stalls. There are always scientific approaches or technologies that show promise but have not yet attracted enough investigator interest to generate unsolicited applications. NOFOs can seed these areas by explicitly inviting proposals and providing startup funding. When announcements stop, these nascent fields often wither before they mature.

Scientific program staff lose the ability to steward their fields. Part of my job was watching for gaps, talking to researchers about unmet needs, and working with my institute to direct resources toward those problems. That function is being eliminated. What remains is passive grant processing: review whatever applications arrive and fund the highest-scoring proposals. That approach works for well-established research areas with active investigator communities. It fails for everything else.

Why This Matters Beyond NIH

The collapse of NIH funding announcements is part of a larger pattern I have documented in previous essays: restructuring the agency without congressional authorization.

Congress rejected proposals to consolidate NIH’s 27 institutes and centers. But if all institutes are restricted to processing generic unsolicited applications through the same centralized approval system, the functional distinction between them disappears. Why maintain 27 separate entities if none of them can independently set research priorities or direct resources toward identified gaps?

The NOFO collapse accomplishes administratively what could not be achieved legislatively. It strips institutes of the autonomy that made them distinct. It centralizes decision-making under political appointees. It eliminates the scientific stewardship function that program staff have exercised for decades. And it does all of this without a single congressional hearing or recorded vote.

This represents a redefinition of what NIH does. The agency is being transformed from an institution that actively identifies and addresses research needs into a passive funding mechanism that distributes money to whatever proposals happen to arrive. That transformation has profound implications for every disease that depends on NIH research, every population whose health needs are not currently being addressed by unsolicited grant applications, and every future health crisis that will require rapid, coordinated research response.

What Scientific Stewardship Looks Like

I spent 22 years identifying research gaps and writing announcements to fill them. When a new disease emerged, when a population was being overlooked, when a promising area needed resources, we could act. That is what scientific stewardship looks like. It is not perfect. It makes mistakes. It can be slow. But it represents accumulated expertise about what research is needed and how to direct resources toward it.

The data shows that function being eliminated in real time. Active direction of resources toward need has been replaced by passive waiting for whatever applications arrive. Scientific program staff exercising judgment about research priorities have been replaced by political appointees controlling every announcement.

The 151 announcements sitting in forecast limbo for 2025, and the 47 blocked so far in 2026, prove this is not happening because of lack of scientific need or budget constraints. The announcements were written. The problems were identified. The resources were allocated. What changed was who gets to decide whether those announcements become active funding opportunities.

This is not efficiency. This is not streamlining. This is the systematic elimination of scientific stewardship at the world’s largest biomedical research funder.

And most people have no idea it is happening.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources

National Institutes of Health. (2025). Updates to finding NIH funding opportunities and information. https://grants.nih.gov/policy-and-compliance/implementation-of-new-initiatives-and-policies/updates-to-finding-nih-funding-opportunities-and-information

Kaiser, J. (2026, March 3). Delays in awards and funding calls worry NIH-funded researchers. Science. https://www.science.org/content/article/delays-grant-awards-and-funding-calls-worry-nih-researchers

NOFO data was compiled by the author from NIH Guide for Grants and Contracts historical archives and grants.gov records, downloaded March 15, 2026. Analysis covers all funding announcements published from 2012 through March 15, 2026.

(I used Gemini Nano Banana 2 to create this fictional image.)

Multiple proposals to restructure the National Institutes of Health have been debated over the past two years. The White House’s fiscal year 2026 budget proposed consolidating NIH’s 27 Institutes and Centers into eight. A congressional report suggested reducing them to 15. Think tanks floated regional models and block grants to states.

Congress rejected them all. The votes weren’t there. The political support didn’t materialize. The proposals failed to advance through the legislative process that major institutional changes to federal agencies traditionally require.

The restructuring is happening anyway. It proceeds through administrative action, executive authority, and bureaucratic maneuvering that bypasses congressional oversight entirely.

This is not routine executive discretion over agency operations. This is systematic dismantling of an agency’s structure, mission, and capacity without the legislative approval that such changes have historically required. And it is accelerating.

The Proposals Congress Rejected

The most radical proposal came from the Heritage Foundation’s Project 2025: replace much of NIH’s competitive grant-making authority with block grants to states. Federal funds would be distributed to states, which would then allocate research dollars according to their own priorities and processes. This would fragment the national coordination that makes large-scale biomedical research possible.

The White House’s budget proposed consolidating the 27 Institutes and Centers (ICOs) into eight, eliminating several units entirely including the National Institutes for Nursing Research, National Center for Complementary and Integrative Health, Fogarty International Center, and National Institute on Minority Health and Health Disparities. A House Energy and Commerce Committee report suggested a more modest consolidation to 15 institutes. Another blueprint recommended replacing the national structure with four regional entities (Northeast, South, Midwest, and West), each with its own director.

These proposals varied in specifics but shared common elements: fewer institutes, less autonomy for scientific program staff, more centralized political control, and reduction of the coordination mechanisms that enable research at national scale.

None secured congressional support. The most recent budget negotiations resulted in full funding for all 27 institutes, centers and offices. No consolidation language made it into appropriations bills. The legislative branch declined to authorize the restructuring that executive branch officials and their think tank allies had proposed.

That should have settled the matter. It didn’t.

Restructuring Without Votes

When these proposals failed to advance legislatively, a different strategy emerged: implement the restructuring administratively. Offices could be closed without congressional hearings. Review panels could be eliminated through internal reorganization. Funding mechanisms could be centralized through executive policy changes. Institute autonomy could be stripped through bureaucratic directives. Approved budgets could be restricted through executive discretion.

No legislation required. No floor debate. No recorded votes. Just administrative action, presented as routine management decisions.

The following examples illustrate how major institutional restructuring is being accomplished at NIH without the congressional approval it would traditionally require.

Closing Offices Without Hearings

In March 2025, seven employees who had worked at NIH’s Sexual & Gender Minority Research Office were placed on administrative leave without warning or explanation. The office, established in 2015, was created to coordinate research and activities for SGM populations across the NIH. It focused on health disparities in LGBTQ populations: higher rates of certain cancers, mental health challenges, and barriers to care. Its creation was prompted by a 2011 Institute of Medicine (IOM) report and subsequently strengthened by the 21st Century Cures Act signed into law December 2016, which officially mandated the inclusion of SGM populations in NIH research.

In December 2024, all seven staff were preemptively reassigned to other offices in anticipation that the incoming administration would dismantle their unit. They were given little choice in the reassignment. Three months later, they were placed on indefinite administrative leave. The research portfolio they built (peer-reviewed grants, multi-year studies, collaborative networks) was left in limbo.

No formal announcement explained the decision. The office simply stopped functioning. No congressional committee held hearings. No appropriations language authorized the closure. It happened through administrative action alone.

Eliminating Specialized Review Panels

The NIH’s Center for Scientific Review (CSR) had historically reviewed 78% of grant applications, about 66,000 per year. The remaining 22% were reviewed by panels run by NIH’s individual institutes and centers, which brought specialized expertise to particular research areas.

In 2025, the administration eliminated all institute- and center-run review panels. Many of the scientists overseeing them were laid off. NIH leadership framed the change as improving efficiency by consolidating review. But researchers and remaining NIH staff understood it differently: as centralization of control that removes specialized scientific judgment from funding decisions.

When review authority is consolidated, it becomes easier to impose uniform priorities from above. Specialized panels had provided a buffer. They brought deep expertise to narrow fields and maintained some autonomy for institutes to steward their scientific domains. That buffer is gone.

Congress was not consulted. No legislation authorized the elimination of these panels. The restructuring of NIH’s peer review system (a core function of the agency) happened through administrative directive.

Redirecting Institute Missions

The National Institute of Allergy and Infectious Diseases (NIAID), historically central to infectious disease research and pandemic preparedness, has undergone significant reorientation without congressional input.

Staff were directed to remove references to “biodefense” and “pandemic preparedness” from institute materials. An infectious disease research group that had participated in outbreak modeling and early-detection collaborations for years was advised to pivot toward projects with more immediate clinical endpoints. Long-term surveillance modeling, once supported under preparedness funding lines, was no longer prioritized.

When researchers asked how the networks and analytic tools they had built over years of federal investment would be sustained, the answer was uncertain. They were encouraged to explore “alternative partnerships,” which, in many cases, do not exist. Alternative partnerships don’t fund decade-long surveillance infrastructure or maintain expertise during quiet periods between outbreaks.

Preparedness research rarely produces visible returns in quiet years. Its value becomes evident only when a crisis emerges. When such work is deprioritized administratively, institutional memory and technical capacity dissipate gradually. By the time the next pandemic arrives, the networks will be gone. The expertise will have moved to other fields. The surveillance systems will have lapsed.

Congress appropriated NIAID’s budget with the expectation that infectious disease preparedness would remain a priority. The mission redirection is happening without legislative authorization.

Centralizing Control Over Research Priorities

The NIH is eliminating the individual websites of its 27 Institutes, Centers, and Offices. Each one currently maintains a public-facing website explaining its unique scientific mission, research priorities, and funding opportunities. These will be replaced by a single centralized page offering minimal information.

Simultaneously, the institutes are losing authority to issue specialized funding announcements targeted at specific research gaps. The NIH central leadership will now vet all such proposals and has stated its intention to reduce specialized funding opportunities by at least 50%.

This may sound like administrative tidying. It is not. When an institute identifies an emerging scientific need (a gap in antibiotic resistance research, a cluster of unexplained pediatric cancers, a promising but underfunded area of Alzheimer’s biology), it previously had authority to issue a targeted funding announcement, convene expert reviewers, and direct resources toward that gap quickly. This specialized funding mechanism allowed scientific program staff to actively steward their fields rather than passively wait for whatever applications happen to arrive.

Eliminating this capacity serves two purposes. First, it removes scientific decision-making authority from civil servant scientists and concentrates it in political leadership. Second, by erasing the visible distinctions between institutes (their unique missions, their specialized expertise, their targeted priorities), it makes the case for consolidation easier to justify. If all 27 institutes are just processing generic applications through the same centralized review, why maintain 27 separate entities?

The answer is that they weren’t just processing generic applications. They were actively identifying gaps, directing resources toward neglected areas, and exercising scientific judgment about emerging needs. That stewardship function is being systematically dismantled through administrative directive, not legislative reform.

Withholding Congressionally Approved Funding

Perhaps the most brazen example involves funding that Congress has already appropriated and the President has already signed into law.

Weeks after Congress rejected the Trump administration’s proposed cuts to NIH and passed a full-year budget on February 3, 2026, the White House Office of Management and Budget (OMB) has refused to authorize release of the funds. The NIH has not received approval to spend any of the research funding that Congress allocated and the President signed.

Normally, after a budget bill becomes law, agencies automatically receive permission to spend a 30-day portion of their annual funding immediately. This permission is renewed every 30 days until OMB approves the agency’s spending plan, which unlocks the full appropriation.

But the OMB changed those rules. Last August, it revised a document known as Circular A-11 to restrict automatic 30-day portions to cover only essential expenses such as employee salaries. Research grants (the core of NIH’s mission) were excluded from this automatic release.

Until its newly appropriated budget is released, NIH can issue new research awards only using leftover money from stopgap funds approved in November 2025. New grant awards have therefore slowed to a trickle.

A similar situation occurred last year, and the NIH issued about 24% fewer new grants than the ten-year average. This year could be worse than last year if the funding blockade continues.

The mechanism is procedural. The effect is substantive. Congress appropriated the funds. The President signed the law. The money exists. The OMB is simply refusing to release it. This is not budget management. It is nullification of congressional appropriations through executive obstruction.

Compounding this obstruction is the implementation of a new “Unified NIH Funding Strategy” that eliminates paylines as the primary mechanism for grant funding decisions. Under the new policy, NIH staff must now justify each funding decision using multiple, sometimes competing criteria.

All ICOs must ensure their grant funding policies align with NIH’s mission, prioritize scientific merit, consider peer review information in its entirety, integrate topics and approaches relevant to their priorities, promote investigator career development and workforce sustainability, promote broad distribution and geographic balance of funding, consider the total amount and type of NIH funding already available to each investigator, and align with the availability of funds. NIH staff who manage R&D contracts have additional requirements as they must get Presidential appointee approvals for every single action.

Interpreting what these requirements mean individually is challenging. Figuring out how to combine all of these elements into justifiable funding decisions that will satisfy the administration is where the system gums up. NIH staff are navigating this new policy while also managing the funding obstruction, the elimination of specialized review panels, and the loss of institute autonomy over research priorities. The cumulative effect is paralysis masked as process reform.

What’s Being Dismantled

From outside the system, these changes may appear as routine portfolio management: efficiency measures, administrative reorganization, budget discipline. From inside, they represent systematic dismantling of institutional capacity and scientific autonomy without congressional authorization or public debate.

What is being lost is not merely funding or staffing. It is the architecture that makes American biomedical research work at national scale. The coordination mechanisms that pool rare patients across state lines. The specialized review systems that bring deep expertise to narrow fields. The surveillance infrastructure maintained between crises. The institutional autonomy that allows scientific program staff to identify emerging needs and direct resources quickly through targeted funding announcements. The early career training pipelines that build expertise over decades.

These systems were built over 75 years through sustained bipartisan investment and congressional oversight. They function invisibly when they work well. Most people don’t know they exist. That invisibility is precisely what makes them vulnerable to administrative dismantling. By the time the consequences become visible (the missing treatment for a rare disease, the delayed response to an emerging pathogen, the clinical trial that cannot recruit enough patients), it will be difficult to trace the connection back to decisions made in 2025 and 2026.

But the damage will be real. And irreversible. Surveillance networks take years to establish. Clinical trial infrastructure and basic science laboratories require sustained institutional relationships. Expertise, once scattered, does not spontaneously reassemble when political winds shift.

A Precedent That Extends Beyond Science

The restructuring of the NIH through administrative action rather than legislative authorization matters beyond science policy. It establishes a precedent for how any federal agency can be systematically dismantled when an administration cannot secure congressional approval for the changes it seeks.

The pattern is clear: When legislation fails, proceed administratively. Close offices through personnel actions. Eliminate review structures through internal reorganization. Redirect missions through policy directives. Strip institutional autonomy through bureaucratic centralization. Withhold appropriated funds through procedural obstruction. Frame it all as routine management.

This approach threatens any federal agency whose mission or structure protects functions that are politically inconvenient or ideologically disfavored. The NIH is not unique in this vulnerability. It is simply a current target. The environmental agencies, the regulatory bodies, the research institutions, the oversight mechanisms. All are vulnerable to the same pattern of administrative dismantling when legislative authorization cannot be secured.

The question is not whether one supports or opposes any particular restructuring of the NIH. The question is whether major institutional changes to federal agencies should require congressional approval or can be accomplished through unilateral executive action.

Historically, the answer has been clear. Major reorganizations require legislation. That constraint exists not to preserve bureaucratic inertia but to ensure that changes to agencies funded by taxpayers and serving public missions are debated, authorized, and reversible through democratic processes.

When that constraint is bypassed (when Congress says no and the restructuring happens anyway), we are in different institutional territory. The implications extend far beyond any single agency.

Breaking the Silence

In my previous essay, I wrote about why NIH staff cannot speak publicly about what is being lost. They remain silent not from indifference but from risk management. Speaking carries consequences, and staying inside is the only way many believe they can still protect what remains.

This essay is about what they would say if they could.

They would say that Congress rejected these restructuring proposals for reasons: concerns about fragmentation, loss of coordination, elimination of specialized expertise, risk to pandemic preparedness. They would say that implementing the restructuring administratively after legislative rejection is not routine executive discretion. They would say that the coordination mechanisms being dismantled cannot be quickly rebuilt when political conditions change.

They would say that institutional architecture is invisible when it functions well, which is why it can be dismantled without immediate public outcry. And they would say that by the time the damage becomes undeniable (by the time the rare disease goes unstudied, the pandemic spreads unchecked, the clinical trial cannot recruit patients, the cell biology laboratory shuts down), the connection to administrative decisions made without congressional authorization will be difficult to trace.

But they cannot say these things without risking their positions and their ability to protect what institutional capacity still remains. That silence is what I am trying to break. Because what is happening to the NIH matters not only for biomedical research but for the principle that major changes to federal agencies require congressional authorization rather than unilateral executive action.

Congress rejected these proposals. They are happening anyway. That is the story that needs to be told.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources

The Heritage Foundation. (2025). Mandate for Leadership: The Conservative Promise (Project 2025). https://static.heritage.org/project2025/2025_MandateForLeadership_FULL.pdf

House Committee on Energy and Commerce. (2025, June). Reforming the National Institutes of Health: Framework for Discussion. https://d1dth6e84htgma.cloudfront.net/NIH_Reform_Report_f6bbdca821.pdf

Kulldorff, M. (2025, March 26). A Blueprint for NIH Reform. RCJL Journal of the Academy of Public Health. https://doi.org/10.70542/rcj-japh-art-1prfdvs

Gaffney, T. (2025, March 6). NIH puts former Sexual & Gender Minority Office employees on administrative leave. STAT. https://www.statnews.com/2025/03/06/nih-puts-former-sexual-and-gender-minority-office-employees-on-administrative-leave/

Kaiser, J. (2025, March 7). NIH will eliminate many peer-review panels and lay off some scientists overseeing them. Science. https://www.science.org/content/article/nih-will-eliminate-many-peer-review-panels-and-lay-some-scientists-overseeing-them

Taubenberger, J. K., Powers, J. H., & Bhattacharya, J. (2026). The new vision from the National Institute of Allergy and Infectious Diseases (NIAID). Nature Medicine. https://doi.org/10.1038/s41591-025-04160-1

National Institutes of Health. (2026, January 6). NIH institutes, centers, and offices. U.S. Department of Health and Human Services. https://grants.nih.gov/funding/find-a-fit-for-your-research/nih-institutes-centers-offices

Kozlov, M., Witze, A., & Garisto, D. (2026, February 27). White House stalls release of approved US science budgets. Nature. https://www.nature.com/articles/d41586-026-00601-0

National Institutes of Health. (2025, November 21). Implementing a unified NIH funding strategy to guide consistent and clearer award decisions. https://grants.nih.gov/news-events/nih-extramural-nexus-news/2025/11/implementing-a-unified-nih-funding-strategy-to-guide-consistent-and-clearer-award-decisions

The people who understand most clearly what is being lost at the NIH are also the people least able to say so. They are not silent because they are indifferent. They are silent because speaking carries real risk, and remaining inside is the only way they believe they can still protect what remains.

This essay is about what that constraint means, why it matters, and why external voices like mine are not optional commentary but a functional necessity when institutions can no longer speak honestly about their own condition.

I write from outside the NIH, which gives me freedom many current staff do not have. That distance is not a claim to special insight. It is an acknowledgment of asymmetry. For 22 years I worked inside the institution, and I understood then, as I do now, that federal science depends in part on civil servants who are careful with words because they understand the power those words carry. That norm still matters. But the conditions that made it workable are breaking down.

Constraint Is Not the Same as Consent

It is tempting to interpret public quiet as acceptance. If concerns were truly serious, the thinking goes, surely someone inside would say so. But that logic misunderstands how federal agencies behave under pressure.

Civil servants today are operating in an environment where formal whistleblower protections exist on paper, but recent experience has shown that speaking out can still result in retaliation, administrative leave, or termination. Highly visible cases have sent a clear signal: raising concerns publicly may carry personal costs, even when those concerns are grounded in constitutional free speech rights and mission protection rather than politics.

In that context, silence is not endorsement. It is risk management.

Many NIH staff are making a deliberate choice to remain in their roles precisely because they believe their presence still matters. As I described in an earlier essay about the loss of experienced personnel, capacity does not disappear all at once. It thins. It frays. And those who remain often see their primary responsibility as preventing further unraveling.

They are trying to keep grant pipelines moving, preserve scientific review integrity, mentor newer staff, and hold institutional memory in place. They cannot do that if they are removed. So, they absorb strain quietly, hoping to stabilize systems from within.

That choice deserves to be understood for what it is: a laudable effort to protect capacity, not a failure of courage.

A View from Inside the Constraint

Consider a situation that is increasingly common across the NIH.

Even as Congress has explicitly rejected proposals to eliminate or consolidate Institutes and Centers in the most recent funding bill, staff inside the agency remain deeply worried about reorganization. Not because it is congressionally authorized, but because the signals are mixed, the timelines unclear, and the political leadership still desires future restructuring. As of February 2026, 16 of the 27 NIH institutes and centers are operating without permanent directors, following a series of firings and retirements. Several current and former staffers have told reporters that they worry the agency will become more politicized depending on who Health Secretary RFK Jr. approves to fill those open positions.

Inside offices, this uncertainty has concrete effects. Program staff and administrators are spending time documenting their value, mapping functions, and preparing quiet justifications for why their units should continue to exist. Conversations that once focused on scientific opportunity now include contingency planning for institutional survival.

I hear from current NIH staff regularly. They describe a particular kind of exhaustion: not just from overwork, but from the cognitive load of performing normalcy while privately managing fear. Unseen by the public, NIH leaders are quietly deciding which offices to keep and which to merge or eliminate, and there is infighting about which offices should be considered “essential” by the administration. These decisions are not being driven by scientific need. They are being driven by politics.

No one is speaking publicly about this. They cannot. Drawing attention to an office can make it a target. Remaining quiet feels safer than being visible.

This is not paranoia. It is rational behavior in an environment where decisions feel decoupled from formal guardrails, and where being noticed can carry risk.

From the outside, the NIH appears to be functioning. Grants are still reviewed. Awards are still made. This outward continuity is often cited as evidence that concerns are overblown. But internally, people are operating under a persistent background fear that structures they rely on may disappear despite formal assurances to the contrary.

This is what constraint looks like inside a complex institution. And it is closely related to what I previously described as normalization: the ability of systems to continue operating just well enough that deeper erosion remains publicly invisible.

The New Tools of Institutional Pressure

Two recent policy changes illustrate how the pressure on NIH staff is becoming more structural, and why the silence is likely to deepen.

The first is the finalization of the Schedule F rule. The Trump administration has moved forward with a personnel policy that could strip civil service protections from thousands of federal workers. The Office of Personnel Management estimates the rule would affect 50,000 positions across the federal workforce. For the NIH, the concern is specific: program officers and other grant-making officials could be reclassified as Schedule F employees, making them significantly easier to dismiss and far more vulnerable to political pressure over their grant decisions. The rule explicitly states that such workers can be fired for “subverting Presidential directives.”

Jenna Norton, an NIH program officer who has spoken publicly in her personal capacity about conditions at the agency, put it plainly: if program officers were moved under Schedule F, “it will become even harder for us to push back on, you know, concerns internally, it will be harder to blow the whistle on concerns externally, or to use our free speech rights to raise awareness around issues.” As of this writing, no NIH scientific staff, including Norton, have been reclassified under Schedule F. The policy is expected to take effect around March 8, 2026. And yet the consequences of speaking out arrived well before any new rule. Norton was placed on administrative leave in November 2025 simply for speaking out about the administration’s disruptions to the agency. She filed a whistleblower complaint in February 2026 alleging the move was unlawful. Her case illustrates precisely the dynamic this essay describes: the people most knowledgeable about what is happening inside the institution are the ones most exposed to consequences for saying so. Schedule F would not create that dynamic. It would entrench it.

The second pressure is less dramatic but equally corrosive. NIH staff are now required to connect at least one of their 2026 performance goals to the President’s Management Agenda, a federal website whose stated priorities include “eliminating woke, weaponization, and waste,” “downsizing the federal workforce,” and ensuring that “grants go only to high-performing recipients to advance America First priorities.”

This requirement asks scientists whose careers have been built on peer-reviewed, mission-driven biomedical research to formally align their professional goals with a political agenda. It does not require them to say anything publicly. It simply asks them, in writing, as part of the official performance review process, to connect their work to the administration’s ideological priorities.

Taken together, these two policies create a structure in which dissent becomes professionally costly, alignment becomes professionally necessary, and the space for honest scientific judgment quietly contracts. The silence deepens not because people stop caring, but because the institutional incentives for speaking have been systematically removed.

When Voice Is Constrained, It Moves

In periods of stability, most institutional truth-telling happens internally. Problems are surfaced through meetings, advisory councils, review processes, and professional disagreement. External critique plays a secondary role because internal mechanisms still function.

But when those mechanisms weaken, the balance shifts.

The NIH now faces a convergence of pressures: staffing losses, leadership turnover, administrative bottlenecks, Schedule F reclassification, performance goals tied to political priorities, and heightened political scrutiny of grant decisions. At the same time, the institution’s ability to name its own problems publicly has narrowed. The result is a paradox. The need for honest assessment increases just as the cost of candor rises.

This is where external voices enter—not as substitutes for internal governance, but as extensions of it.

Former staff, retirees, and outside experts often carry deep institutional knowledge without carrying current institutional risk. When they speak, they are not bypassing the NIH. They are absorbing pressure that insiders cannot safely take on. This is not whistle-blowing in the cinematic sense. It is a form of institutional relay, moving observations outward so they can be processed without immediate retaliation.

Importantly, this kind of external critique is often more conservative than it appears. It does not seek to de-legitimize public science or undermine trust. It seeks to preserve core functions by naming stresses that cannot be fully articulated from inside the building.

Why Distance Has Become a Form of Protection

The danger arises when external critique is dismissed as alarmist or disloyalty simply because it comes from outside formal authority. That dismissal misunderstands the conditions under which such voices emerge.

At the NIH, the loss of experienced staff, the thinning of scientific review capacity, and the normalization of overload are not abstract trends. They are operational realities that insiders manage daily. But describing their cumulative effects publicly can be interpreted as undermining confidence, even when the intent is to prevent further damage.

When endurance is re-framed as resilience, and silence is mistaken for stability, institutions lose a critical feedback signal. Problems are no longer contested in public view. They are quietly redistributed downward, absorbed by individuals who “make do” until they cannot. External voices disrupt that normalization. They make visible what professionalism and loyalty often keep hidden—not because insiders are failing to speak, but because the system no longer allows them to do so safely.

There is an irony here that is uncomfortable but important. The people speaking most freely about what is happening to American biomedical science are increasingly those who no longer work inside it. That is not a coincidence. Distance creates safety. Safety allows candor. And candor, right now, is the thing institutions most need and least allow.

We should not mistake that distance for disloyalty. In many cases, it is the opposite. When internal voice is systematically constrained, speaking from outside is not abandonment. It is the only form of institutional protection still available. This is not a sustainable substitute for strong internal governance. It is a rational response to a moment when internal voice is constrained and external understanding lags dangerously behind operational reality.

If we care about preserving publicly funded biomedical science, we need to stop treating external critique as a threat and start asking what conditions made it necessary.

What the Silence Is Telling Us

There is a final question worth sitting with.

We have built, over 75 years, one of the most productive biomedical research enterprises in human history. The people who understand it best—who spent decades inside it, who know which scientific review panels are gone, which institutes are being quietly reorganized, which grants are going unfunded not for scientific reasons but for political ones—many of those people cannot say what they know without risking their livelihoods.

That is the signal. Not the official statements. Not the press releases about continued commitment to science. Not the reassurances that Congress passed a full budget, grants are still being reviewed, and awards still being made.

The signal is the silence itself. And what it tells us is that the conditions under which honest institutional self-assessment was possible have changed.

The question is not why some of us can now say these things from the outside. The question is why so many cannot say them from the inside. That silence is not a minor concern. It is a warning about the institution the world has counted on, for 75 years, to turn public investment into lifesaving knowledge.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources

Oza, A. (2026, February 6). Scientists worry finalizing ‘Schedule F’ rule will further politicize NIH grant decisions. STAT. https://www.statnews.com/2026/02/06/nih-grant-reviewers-schedule-f-designation-politicizes-science/

Stolberg, S. G. (2026, February 2). N.I.H. worker who criticized Trump seeks whistle-blower protection. The New York Times. https://www.nytimes.com/2026/02/02/us/politics/nih-worker-trump-whistle-blower-protection.html

Kaiser, J. (2025, November 24). NIH shake-up to grant decision-making sparks concern over political meddling. Science. https://www.science.org/content/article/nih-shake-grant-decision-making-draws-concerns-political-meddling

Bendix, A. (2026, February 15). National Institutes of Health faces leadership vacuum as director positions sit open. NBC News. https://www.nbcnews.com/health/health-news/national-institutes-health-director-positions-unfilled-rcna257834

President’s Management Agenda. (2026). Performance.gov. https://www.performance.gov/pma/

Hersher, M., & Mervis, J. (2026, January 26). U.S. government has lost more than 10,000 STEM Ph.D.s since Trump took office. Science. https://www.science.org/content/article/u-s-government-has-lost-more-10-000-stem-ph-d-s-trump-took-office

In recent months, as National Institutes of Health (NIH) advisory councils have been depleted, thousands of federal scientists have departed, and research capacity has contracted, I’ve heard the same refrain from some voices: These cuts are necessary. NIH has lost its way. Government shouldn’t be funding science anyway.

The most sophisticated version of this argument comes from the Cato Institute, a libertarian think tank. In a February 2026 blog post titled “Restoring the NIH Mission,” authors John F. Early and Terence Kealey argue that NIH funding is not just wasteful but actively harmful. They claim that before Sputnik, when 95% of biomedical research was privately funded, research was “twice as effective” as it is today. They identify 43% of current NIH spending as wasteful or harmful. And they argue that government funding “crowds out” private innovation.

Their argument deserves serious engagement—not because it’s correct, but because it represents the intellectual foundation for the dismantling currently underway at the NIH. It’s also worth engaging because even these libertarian critics find themselves alarmed by some of what the current NIH leadership is proposing.

This essay examines their claims systematically. What it reveals is not a case for eliminating public biomedical science funding, but rather a profound misunderstanding of what knowledge infrastructure requires, how scientific progress actually works, and what happens when we confuse market efficiency with scientific effectiveness.

The Correlation Fallacy

The central claim in the Cato argument rests on a single graph. It shows that NIH spending increased fivefold after Sputnik, while the rate of life expectancy improvement fell by more than half. The implied causation is clear: more NIH spending made biomedical research less effective.

This is a textbook example of mistaking correlation for causation.

Life expectancy gains slowed in all developed countries during this period, regardless of their research funding models. The United Kingdom, France, Germany, and Japan all experienced similar patterns. Why? Because the nature of biomedical research challenges changed fundamentally.

In the early 20th century, life expectancy gains came primarily from addressing acute threats: infectious disease, infant mortality, injuries, and nutritional deficiencies. These were, relatively speaking, low-hanging fruit. Sanitation systems, antibiotics, vaccines, and basic public health interventions produced dramatic results quickly.

By the 1960s, we’d largely solved those problems in developed countries. What remained were fundamentally harder challenges: cancer, heart disease, stroke, Alzheimer’s, diabetes, autoimmune conditions, and rare genetic disorders. These are chronic, complex, multifactorial diseases that take decades to understand and treat.

Comparing life expectancy gains across these two eras is like complaining that running the first mile of a marathon was easier than running the last mile. The slowdown doesn’t indicate failure—it indicates that the problems got harder.

The Cato authors ignore confounding variables: obesity, the opioid epidemic, income inequality, and rising healthcare costs. By their logic, we should also conclude that the internet, color television, and the decline of smoking in public all caused the life expectancy slowdown, since they occurred in the same timeframe.

What this correlation actually shows is something the Cato authors don’t want to acknowledge: that in the absence of easy wins, sustained public investment becomes even more necessary. The diseases that remain are precisely the ones that markets won’t fund research on because the timelines are too long, the risks too high, and the returns too uncertain.

The “Mission Model” Myth

The Cato authors contrast what they call the “mission model” (focused, goal-oriented research) with the “pipeline model” (exploring interesting questions in hopes valuable information will emerge). They claim the pre-Sputnik era’s mission model, dominated by private funding, was more effective.

This argument rests on a romantic myth about how scientific progress actually happens.

The achievements they point to from the early 20th century—antibiotics, vaccines, basic surgical techniques—were indeed often developed with specific applications in mind. But they built on decades of “useless” basic research that had no immediate application when it was conducted.

Penicillin was discovered by accident by Alexander Fleming, a government-funded researcher at St. Mary’s Hospital in London. The germ theory of disease, which made antibiotics conceivable, emerged from basic research on fermentation and microscopy. The structure of DNA, which underlies all of modern genetics and cancer treatment, came from curiosity-driven research with no immediate application.

The Human Genome Project is the perfect example of why the mission vs. pipeline distinction is false. It had a specific goal: sequence the entire human genome. But that goal was only possible because of decades of “pipeline” research on DNA structure, sequencing technology, and computational biology. And the results of the genome project weren’t immediate treatments—they were foundational knowledge that is still yielding new therapies 20 years later.

The Cato authors also ignore that their preferred “mission model” era included massive research failures and dead ends that never made it into the history books. We remember penicillin; we forget the hundreds of compounds that were tested and failed. Survivorship bias makes the past look more efficient than it was.

More fundamentally, the distinction between mission and pipeline research is artificial. All successful biomedical research requires both: curiosity-driven exploration to understand basic mechanisms, and focused application to turn that understanding into treatments. You can’t have one without the other.

The “Crowding Out” Fantasy

A central claim in the Cato argument is that government funding “crowds out” private innovation. The theory is that pharmaceutical companies and private foundations would fund basic biomedical research if only the government would get out of the way.

This is contradicted by the explicit statements of pharmaceutical companies themselves.

Major pharmaceutical companies openly state that they don’t fund basic research because it’s too risky, too slow, and returns are too uncertain. They focus on applied development: taking discoveries from basic research and turning them into drugs. This isn’t a secret—it’s their stated business model.

Consider the examples already discussed in my previous essays:

HIV/AIDS research in the 1980s: Pharmaceutical companies saw no profit in it. The patient population was stigmatized, mostly uninsured, and dying too quickly to generate returns. It was NIH-funded researchers who developed AZT and ran the clinical trials. Private companies only entered the field after the government proved antiretroviral therapy worked.

Antibiotic development today: Pharmaceutical companies are abandoning this field entirely because resistant bacteria kill patients too quickly to recoup development costs. The business model doesn’t work. We’re heading toward a post-antibiotic era not because we lack capability, but because markets won’t fund it.

mRNA technology: This was supported by decades of NIH funding during periods when clinical application was far from guaranteed. When COVID-19 hit, private companies built on that foundation—but they didn’t create it.

The “crowding out” theory assumes private companies were eager to fund this work but stopped because government was doing it. The reality is the opposite: private companies have never been willing to carry the uncertainty of basic research. NIH doesn’t crowd out private research—it creates the foundation that makes private research possible.

This is complementary, not competitive. And it works precisely because each sector does what it does best: government funds long-term, uncertain exploration; private sector develops applications and brings them to market.

What They Call “Waste” Reveals the Ideology

The Cato authors identify 28 categories of NIH research—accounting for 43% of spending—that they claim are “harmful or wasteful.” Examining what they put in this category is revealing, because it exposes the ideological assumptions underlying their efficiency arguments.

They identify four types of waste:

First, they claim NIH funds projects that “reject the scientific method and make claims their own data show are false.” Their primary example is hormone replacement therapy research that they claim falsely linked HRT to breast cancer risk, leading to harmful recommendations.

This criticism fundamentally misunderstands how science works. The Women’s Health Initiative study on HRT was exactly what rigorous science looks like: a large randomized controlled trial that produced unexpected results, triggered debate, led to reanalysis, and ultimately refined our understanding. The fact that initial findings were later nuanced doesn’t mean the research was fraudulent—it means science was working as designed. Knowledge advances through exactly this kind of iterative process.

Second, they claim NIH funds research that “endangers liberty by seeking to prescribe government compulsion of behavior.” Their example: “shaken baby syndrome” research that, they argue, falsely promotes the theory that only abuse can cause certain symptoms, leading to wrongful prosecution of parents.

This is an extraordinary claim. They’re arguing that research on pediatric head trauma—research that helps doctors distinguish between accidental injury and abuse—is government overreach because it might support prosecution of child abusers. The prioritization here is striking: they’re more concerned about “liberty” (the liberty of potential abusers to avoid prosecution?) than about child safety or medical accuracy.

Research on abusive head trauma is legitimate medical science. Courts may misuse scientific findings, but the solution is better scientific communication and legal standards, not eliminating research on child injury.

Third, they claim NIH funds projects that are “irrelevant to health.” They cite two examples: a study on the effects of community choir singing on elderly health, and research examining the relationship between mortgage foreclosures and emergency room visits.

Both of these are actually well-established areas of public health research. Social determinants of health—how housing, income, community engagement, and social isolation affect health outcomes—is scientifically rigorous and has led to important interventions. The choir study examines social engagement and cognitive stimulation in aging populations, which has clear relevance to dementia prevention. The foreclosure study examines stress, housing instability, and health access—all documented health factors.

Dismissing this research as “irrelevant to health” simply because it’s not molecular biology reveals a cramped view of what public health requires. Health isn’t just pharmaceuticals and surgery. It’s also preventing disease through understanding social and environmental factors.

Fourth, they call smallpox research “pure waste” because there have been no cases in 48 years. This demonstrates dangerous naïveté about biosecurity. Smallpox research maintains expertise to manufacture vaccines if needed, defends against bioterrorism, and develops treatments for related poxviruses that do still cause disease. Calling this “waste” is like calling fire insurance waste because your house hasn’t burned down.

What unites all these “waste” categories is not inefficiency but ideology. The Cato authors reject research that:

• Examines social and behavioral factors in health (too collectivist)

• Supports public health interventions (too interventionist)

• Provides evidence for child protection (threatens parental authority)

• Prepares for rare but catastrophic threats (too precautionary)

This isn’t budget analysis. It’s libertarian philosophy dressed up as efficiency review.

The Political Reallocation Scheme

The Cato authors critique what they call “the really bad news”: the current NIH director’s proposal to “spread the wealth to more universities.” But the context is crucial and deeply troubling.

The current NIH Director, Jay Bhattacharya, has proposed redirecting funding away from elite coastal universities—mostly in blue states—to heartland institutions in red states. This isn’t hypothetical. Bhattacharya floated his plan during his February 3, 2026 Senate Health, Education, Labor and Pensions Committee hearing on modernizing the NIH. He explicitly framed this as part of the “Make America Healthy Again” (MAHA) mission and he knows it will garner support from MAGA Republicans on Capitol Hill. Bhattacharya’s plan is to break the current link between research support and facility/overhead funding by forcing institutions to compete separately for each. In his vision, scientists whose research proposals are approved would need to move to states with “inexpensive labs” to conduct their work, creating what he calls “a market for science, funded by the government.”

Let’s be clear about what this means in practice:

Institutions like Harvard, MIT, Stanford, Johns Hopkins, and UMass Chan have spent 60-70 years building biomedical research infrastructure in partnership with NIH. They have specialized equipment, core facilities, biosafety labs, animal facilities, imaging centers, genomics platforms, and computational resources that took decades and billions of dollars to build. They have research hospitals where scientists can move discoveries from bench to bedside. They have trained workforces, institutional knowledge, and collaborative networks.

Bhattacharya’s proposal would defund this infrastructure not because it’s ineffective, but because it’s in the wrong states politically. The justification is geographic equity, but the subtext is nakedly partisan: reward Republican constituencies, punish Democratic ones.

This is not “letting the market decide,” as Bhattacharya claims. Markets don’t fund research infrastructure in Oklahoma because there’s no revenue model for it. This is government picking winners based on political geography rather than scientific merit.

Even former Republican NIH Director Elias Zerhouni, who served under George W. Bush, opposes this plan: “When you pick and choose, the quality goes down. The government is not good at picking and choosing. If the government gets into it, it becomes a political thing, and I hate that sort of politically motivated allocation of federal resources. It should be based on merit.”

The irony is rich: the Cato authors oppose Bhattacharya’s plan because it’s too much government intervention. They want pure market allocation (which would mean essentially no basic research funding at all). Bhattacharya wants political allocation disguised as geographic equity. Neither wants what we currently have: merit-based peer review.

What this reveals is that the real debate is not about efficiency versus waste. It’s about who controls knowledge production and for what ends. The current system, for all its flaws, allocates funding based on scientific merit as judged by peer review. Bhattacharya wants to allocate based on political loyalty. Cato wants to allocate based on market returns. Both would destroy the capacity for long-term, uncertain, basic research that has produced every major biomedical breakthrough of the past 75 years.

What This Debate Actually Reveals

Step back from the specific arguments about correlation, crowding out, and waste categories, and a pattern emerges. These critiques of NIH share a common thread: they all assume that knowledge production should justify itself by criteria external to science.

For the Cato authors, the criterion is market returns. Research should be funded if and only if private investors believe it will generate profit. This sounds reasonable until you remember that markets systematically undervalue long-term, uncertain, and collectively beneficial research. The very features that make basic research valuable—its long-time horizons, its unpredictable applications, its public good nature—are precisely what make it unmarketable.

For Bhattacharya, the criterion is political geography. Research should be funded in ways that reward particular constituencies and particular regions, regardless of where expertise and infrastructure currently exist. This sounds like fairness until you recognize that scientific capacity isn’t fungible. You can’t simply move a cancer research program from Boston to Alabama and expect the same results, any more than you could move Silicon Valley to Iowa and expect the same innovation ecosystem.

What both positions reject is the idea that scientific merit, as judged by scientific peers, should be the primary criterion for allocation. They reject it for different reasons—Cato because they distrust government, Bhattacharya because he distrusts the scientific establishment—but the effect is the same: the dismantling of a system that, whatever its flaws, has been extraordinarily productive.

Consider what we’ve actually gotten from 75 years of NIH funding allocated primarily by peer review:

Vaccines for polio, hepatitis B, and HPV. HIV antiretroviral therapy that turned a death sentence into a manageable chronic condition. The Human Genome Project. mRNA vaccine technology. Heart disease prevention through understanding of factors such as cholesterol, blood pressure, and smoking. Cancer immunotherapies. Gene therapies for sickle cell disease. SSRIs and other mental health interventions. The 56% reduction in heart disease deaths between 1950 and 1996. The 54% reduction in AIDS deaths since 2010.

None of these were guaranteed when the research started. None had obvious market applications. All required sustained public investment over decades. All built on supposedly “wasteful” basic research that explored “irrelevant” questions years before applications became clear.

The Cato authors look at this record and see inefficiency because life expectancy gains slowed. What they miss is that the slowdown reflects the difficulty of the remaining challenges, not the failure of the approach. We’ve solved the easy problems. What remains requires exactly the kind of long-term, uncertain, basic research that markets won’t fund and political reallocation will destroy.

What We’re Actually Arguing About

The debate over NIH funding is often framed as a technical question about efficiency and effectiveness. Should we fund more or less research? Which diseases deserve priority? How do we measure return on investment?

But the Cato Institute piece, combined with Bhattacharya’s political reallocation scheme, reveals that this is fundamentally a debate about institutional control and the purpose of knowledge production.

The question is not “Does NIH waste money?” Every large institution wastes some money. The question is: “Compared to what alternative?”

Compared to market allocation? Markets don’t fund basic research. We know this because when NIH funding declines, private funding doesn’t replace it—basic research simply doesn’t get done. The diseases that lack profitable treatments don’t get studied. The long-term questions don’t get asked.

Compared to political allocation? What would see if funding decisions are driven by geography and partisan loyalty rather than scientific merit? Expertise will be destroyed. Infrastructure built over decades will be abandoned. Research capacity doesn’t shift—it disappears.

The current system of peer review is imperfect. It can be conservative, favoring established researchers and safe projects. It can be slow to recognize new paradigms. It has biases around age, gender, race, and institutional prestige. But these are problems that can be addressed within the system through reforms like anonymous review, funding for early-stage investigators, and evaluation criteria that reward innovation. They don’t justify abandoning peer review altogether in favor of market returns or political cronyism.

What the Cato authors and the current NIH leadership share is a conviction that the scientific establishment cannot be trusted to allocate resources wisely. This conviction leads them to very different prescriptions—markets versus political control—but it rests on the same foundational distrust of scientific self-governance.

This distrust is not entirely unfounded. The scientific establishment has made mistakes. It can be resistant to change. It sometimes protects its own interests over the public good.

But the alternative—letting markets or politicians decide which questions get asked and which diseases get studied—is far worse. Because markets optimize for profit, not health. And politicians optimize for self-interest, not knowledge.

Science, for all its flaws, optimizes for understanding. And over the long term, understanding is what produces breakthroughs.

What Public Science Actually Requires

The arguments against NIH funding, whether from libertarian think tanks or political appointees, rest on a fundamental misunderstanding of what knowledge infrastructure requires.

They assume that scientific research can be evaluated like other government programs: by immediate outputs, by cost-benefit ratios, by geographic distribution, by market demand.

But knowledge production doesn’t work that way.

Breakthrough biomedical discoveries don’t arrive on schedule. They emerge from decades of accumulated understanding, built by hundreds of researchers asking questions whose value only becomes clear in retrospect. You can’t predict which line of inquiry will matter. You can’t optimize it like a supply chain. You can’t move it to wherever land is cheapest.

This doesn’t mean NIH funding should be unlimited or unaccountable. It means that accountability requires different metrics than we use for other government programs.

The right question is not “Did this grant produce a treatment this year?” but “Did this grant advance understanding in ways that enable future biomedical discoveries?”

The right question is not “Could the private sector fund this research?” but “Is anyone else positioned to ask these questions at all?”

The right question is not “Does this research fit our political priorities?” but “Does this research address important questions using rigorous methods?”

When 77% of Americans say they trust scientists, what they’re expressing is an intuition that scientific knowledge matters. But trust without institutional support is meaningless. Trust without sustained funding for basic discovery is performative. Trust that allows systematic dismantling of research capacity is betrayal.

If we want the next generation of breakthroughs—treatments for Alzheimer’s, cures for rare cancers, vaccines for future pandemics, solutions to antibiotic resistance—we need to defend the institutions that make those breakthroughs possible.

That means rejecting the false choice between market efficiency and government waste. It means insisting that some things—like understanding how the human body works and what makes it fail—are too important to be left to quarterly earnings reports or political patronage.

It means recognizing that the slowdown in life expectancy gains is not evidence that NIH is failing. It’s evidence that the problems we face now are harder than the problems we’ve already solved. And that makes sustained public investment more necessary, not less.

The Cato Institute wants us to believe that 95% private funding worked better in 1950. What they don’t acknowledge is that 1950’s diseases were fundamentally simpler than 2026’s diseases. That the infrastructure we’ve built over 75 years cannot be recreated by market forces alone. That knowledge, once lost, doesn’t spontaneously return when conditions improve.

The current NIH leadership wants us to believe that moving research from Massachusetts to Alabama will make science fairer. What they don’t acknowledge is that expertise and infrastructure take decades to build and can be destroyed in months. That geographic equity purchased by destroying scientific capacity helps no one.

What both positions offer is not efficiency but abandonment. Not reform but demolition.

The question before us is not whether NIH is perfect. It’s whether we’re willing to defend imperfect institutions that produce irreplaceable knowledge, or whether we’ll let them be dismantled in the name of ideological purity.

I spent 22 years at NIH. I know its flaws intimately. But I also know what it makes possible. And I know that once this capacity is gone, we won’t get it back.

The choice is ours.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources

Early, J. F., & Kealey, T. (2026, February 6). Restoring the NIH mission: Some good news, some not-so-good news, and some really bad news. Cato Institute. https://www.cato.org/blog/restoring-nih-mission-some-good-news-some-not-so-good-news-some-really-bad-news

Office for National Statistics. (2018, August 6). Changing trends in mortality: An international comparison: 2000 to 2016. https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/lifeexpectancies/articles/changingtrendsinmortalityaninternationalcomparison/2000to2016

Dieleman, J. L., et al. (2018). The slowing pace of life expectancy gains since 1950. BMC Public Health, 18(1), 151. https://doi.org/10.1186/s12889-018-5058-9

Woolf, S. H. (2023). Falling behind: The growing gap in life expectancy between the United States and other countries, 1933–2021. American Journal of Public Health, 113(9), 970–980. https://doi.org/10.2105/AJPH.2023.307310

National Research Council. (2010). Current model for financing drug development. In Advancing Regulatory Science for Medical Countermeasure Development. National Academies Press. https://www.ncbi.nlm.nih.gov/books/NBK50972/

Avorn, J. (2025, October 22). Pharmaceutical industry money can’t replace NIH funding. STAT. https://www.statnews.com/2025/10/22/nih-budget-cuts-pharmaceutical-industry-research/

Cleary, E. G., Jackson, M. J., Zhou, E. W., & Ledley, F. D. (2023). Comparison of research spending on new drug approvals by the National Institutes of Health vs the pharmaceutical industry, 2010-2019. JAMA Health Forum, 4(4), e230511. https://doi.org/10.1001/jamahealthforum.2023.0511

National Research Council. (2011). The impact of publicly funded biomedical and health research. In Measuring the Impacts of Federal Investments in Research. National Academies Press. https://www.ncbi.nlm.nih.gov/books/NBK83123/

Kupferschmidt, K. (2020, September 9). Why big pharma has abandoned antibiotics. Nature. https://www.nature.com/articles/d41586-020-02884-3

UNAIDS. (2025). Global HIV & AIDS statistics—Fact sheet. https://www.unaids.org/en/resources/fact-sheet

Dzau, V. J., & Yamamoto, K. R. (2025). Government-funded health and biomedical research is irreplaceable. NAM Perspectives. National Academy of Medicine. https://doi.org/10.31478/202508b

Lim, E. (2026, February 3). Democrats just handed RFK Jr. billions more than he asked for. It was a big risk. POLITICO. https://www.politico.com/news/2026/02/03/democrats-handed-rfk-jr-billions-more-than-he-asked-for-00763792

Freyer, F. J. (2026, February 4). Congress nixes Trump research funding cuts, easing fears among Mass. scientists and lawmakers. The Boston Globe. https://www.bostonglobe.com/2026/02/04/metro/congress-rebukes-trump-research-funding-cuts/

Kennedy, B., & Kikuchi, E. (2026, January 15). Americans’ confidence in scientists. Pew Research Center. https://www.pewresearch.org/science/2026/01/15/americans-confidence-in-scientists/

In January 2026, while 77% of Americans told Pew Research they trusted scientists, the majority of NIH’s 25 advisory councils were operating with less than half their full complement of members. These councils, required by law to approve new grants, have been systematically depleted through expired terms, withdrawn nominations, and unfilled vacancies. At the same time, 15 of the 27 NIH institutes and centers are operating without permanent leadership, following a series of firings and retirements. No press conference announced this governance crisis. No emergency session was convened. The councils simply kept shrinking, and leadership positions remained vacant. And most Americans, the same ones who say they value science, have no idea that this dismantling of the NIH is happening.

This is the paradox at the heart of science policy today. Poll after poll shows solid majorities saying they trust scientists to act in the public’s best interests. In the latest Pew Research Center survey, 77 percent of U.S. adults say they have at least a fair amount of confidence in scientists, a figure that remains strong compared with the public’s trust in many other institutions. And internationally, large-scale studies of tens of thousands of people across dozens of countries find that most people think scientists are qualified and that scientific methods are the best way to determine truth.

And yet, that broad support has not translated into structural protection for the institutions on which science depends. In prior essays I have written about staff losses, eroded governance, shrinking capacity, and normalized contraction not because science is unpopular, but because confidence and power are not the same thing. A large share of the public can value science in the abstract while political actors undermine the very infrastructure that produces it.

This essay is about that paradox: how science can be widely valued and still be vulnerable. How popular approval can coexist with institutional erosion. How societies with high trust in science, or at least in scientists, can nonetheless allow governance practices, administrative bottlenecks, and shifting political priorities to hollow out research capacity without triggering sustained public pressure to repair it.

Understanding this gap between support and leverage is not discouraging. It’s clarifying. It helps explain why people on all sides of the political spectrum can say they believe in science even as they accept narratives that justify doing less science, or as they shrug at changes that dismantle core structures. And it sets the stage for asking a deeper question: What would it take for public support to actually matter in the way most people assume it does?

The American Story We Keep Telling Ourselves

Part of what makes this moment so disorienting is that it collides with a deeply American story about how progress is supposed to work.

In the United States, we are taught to prize individual initiative, self-reliance, and market-driven solutions. Government, by contrast, is often framed as a necessary but suspect actor: something that should be kept small, lean, and constantly justified. Public spending is treated less as collective investment and more as a cost that must defend its own existence.

This worldview has a long lineage. It draws from libertarian ideas about freedom and efficiency, from frontier myths about rugged independence, and from decades of political rhetoric that equates public institutions with waste and private enterprise with virtue. Over time, with lots of repetition and persuasive marketing by corporate interests, these ideas have become pervasive.

Science does not sit comfortably inside this story.

Publicly funded science is slow, collective, and uncertain by design. It depends on institutions, continuity, and shared infrastructure. Its benefits often emerge over decades and cannot be traced to a single product, company, or individual. In other words, it looks exactly like the kind of thing American political culture has been persuaded to mistrust.

This worldview didn’t emerge naturally. It was cultivated. In the 1970s and 80s, think tanks like the Heritage Foundation and the Cato Institute explicitly framed government-funded science as ‘picking winners and losers.’ The same pattern appeared with climate science, where ExxonMobil’s own researchers confirmed global warming in the 1970s, yet the company spent decades funding campaigns to discredit public climate research as ‘government overreach.’ The strategy wasn’t to make people hate science. It was to make them believe government shouldn’t be doing it.

Why This Framing Is So Seductive

The appeal of the libertarian worldview is not hard to understand. It promises freedom from bureaucracy. It flatters individual effort. It suggests that excellence will naturally rise if constraints are removed.

But applied to science, this logic quietly distorts reality.

Scientific progress does not emerge from isolated brilliance alone. It requires stable funding mechanisms, peer review systems, regulatory oversight, training pipelines, data infrastructure, and institutional memory. None of these are visible to the public in the way a breakthrough discovery is. All of them look, from the outside, like overhead.

When the value of science is reduced to outcomes alone, the structures that make those outcomes possible begin to seem expendable. The system is asked to behave as if it were a collection of independent actors rather than an ecosystem.

This is where libertarian instincts and goalpost shifting converge. If science is imagined as something that should function like a market, then fewer grants, fewer staff, and leaner institutions are not warning signs. They are proof of discipline.

The fact that people are ‘making do’ becomes evidence that the cuts were justified.

Why Markets Cannot Carry Science on Their Own

This faith in market-driven solutions is not evenly distributed. Partisan divides on science are evident in the Pew poll when pollsters asked whether the private sector will ensure enough progress is made. While 54 percent of Republicans said yes, only 21 percent of Democrats did so. Confidence that the private sector will adequately fund science reflects a fundamental misunderstanding of research ecosystem incentives.

Private funding plays an important role in some research, but it follows a different logic. It is most available when discoveries can be monetized, timelines are predictable, and returns are foreseeable. That model works well for applied development. It does not accommodate basic science.

Basic biomedical research is slow, risky, and full of dead ends. It often takes decades before its value becomes clear, and even longer before benefits reach patients. That is precisely why, since World War II, the U.S. government has played a central role in funding it. Not because markets failed, but because private sector markets were never designed to carry this kind of uncertainty.

The U.S. strategy of publicly funding biomedical research has paid off repeatedly. Over the past half-century, this investment has yielded enormous dividends for health and medicine.

Consider what didn’t happen when markets led. In the 1980s, pharmaceutical companies saw no profit in HIV/AIDS research. The patient population was stigmatized, mostly uninsured, and dying too quickly to generate returns. It was NIH-funded researchers at the National Cancer Institute who developed AZT as the first antiretroviral therapy. It was government scientists who ran the clinical trials no company would touch.

At their peak around 2004-2005, annual AIDS deaths reached approximately 2 million globally. By 2010, this had declined to about 1.4 million. Today, that number is under 630,000, representing a 54% decline from the 2010 figure. Every person alive today because of antiretroviral therapy owes that life to research that had no business case, no quarterly earnings potential, and no private investor willing to wait the 15 years it took to work.

Or consider what’s happening right now. Pharmaceutical companies are abandoning antibiotic development because resistant bacteria kill patients too quickly to recoup drug development costs. The business model doesn’t work. So, we’re heading toward a post-antibiotic era not because we lack scientific capability, but because we’ve accepted a framework where knowledge production must justify itself to shareholders.

The Human Genome Project is another canonical example. It was not a commercially viable undertaking when it began. The timelines were too long, the applications too speculative, and the costs too high. Yet the public investment created a foundation that now underlies advances in cancer treatment, rare disease diagnosis, and gene-based therapies, including recent breakthroughs for sickle cell disease.

The same is true for mRNA technology. Long before it became central to the COVID-19 response, it was supported through sustained public funding, much of it during periods when clinical application was far from guaranteed. When the pandemic arrived, that accumulated public investment made rapid vaccine development possible.

Other successes include developing vaccines for polio, hepatitis B, and HPV; neurological discoveries supporting the development of SSRIs and other mental health interventions; heart disease prevention through a new understanding of the dangers of high blood pressure, cholesterol, and smoking, which contributed to a 56 percent decline in heart disease deaths between 1950 and 1996; new treatments for breast, lung, prostate, and childhood cancers as well as new immunotherapies; and prevention of Sudden Infant Death Syndrome through the ‘Back to Sleep’ campaign.

None of these outcomes were accidents. They were the result of a robust publicly funded biomedical research system willing to invest in knowledge long before its uses were obvious.

What Individualism Obscures

An individualistic, market-centered lens struggles to see these dynamics. It values efficiency over endurance and outcomes over infrastructure. It asks what science produces, but not what science requires.

When public investment is evaluated primarily through the lens of short-term returns, the structures that make long-term discovery possible begin to look expendable. People simply may not appreciate that the life-saving successes mentioned above took decades to achieve. This may create a false sense that medical advances simply happen on their own given enough time. Publicly funding biomedical research comes to feel indulgent. Maintaining institutional capacity appears excessive. Staff, review systems, and governance mechanisms are treated as overhead rather than as the scaffolding of discovery.

This is where cultural assumptions quietly shape policy outcomes.

If government is presumed to be inherently bloated, then reducing it feels virtuous, even when what is being reduced is scientific capacity itself. If success is measured by doing more with less, then endurance becomes proof that cuts were justified.

This is how goalposts move without formal debate.

The consequences of this cultural framing are not abstract. Here’s what this looks like in practice: A postdoctoral researcher spends three years developing expertise in a rare disease mechanism. Their lab loses NIH funding. They leave science for industry, not because they wanted to, but because they have student loans and a family. That specific knowledge, those three years of accumulated insight, simply vanishes from the field.

Six years later, a patient presents with symptoms no one can explain. The answer was in that researcher’s unpublished work, the experiments that were halfway done when the funding stopped. But that knowledge is gone. The researcher is now optimizing ad algorithms. The patient remains undiagnosed.

This is not a hypothetical. Between 2003 and 2015, the NIH’s purchasing power declined by more than 20% in real dollars. During that same period, the number of PhD scientists leaving research careers reached historic highs because the system lost capacity to employ them.

The loss of STEM PhDs is accelerating under the Trump administration. A new Science analysis of White House Office of Personnel Management data reveals a dramatic surge in the number of employees with a PhD who left employment at federal research agencies in 2025. At the 14 agencies that Science examined, PhD departures outnumbered new hires in 2025 by a ratio of 11 to one, resulting in a net loss of 4,224 STEM PhDs. Losses were steepest at the National Institutes of Health and the National Science Foundation. The most common reasons for departures were retirements and quitting, suggesting that the lure of buyout offers, or disagreement with Trump policies, likely influenced many decisions to leave.

We aren’t just losing studies and labs. We are losing a generation of basic scientific expertise that cannot be recovered by simply restoring budget lines.

How Narrative Becomes Policy

The cultural story about government inefficiency and market efficiency doesn’t just shape what people believe. It shapes what politicians can propose without backlash.

When the OMB proposes cutting NIH funding, the justification isn’t that science is worthless. It’s that NIH needs to ‘streamline’ and ‘eliminate redundancy,’ language borrowed directly from corporate restructuring. The assumption, unstated but powerful, is that science has gotten bloated and needs discipline.

No one asks what will happen to the research in progress. No one demands to know which diseases will go unstudied, which training programs will close, which international collaborations will end. The frame makes those questions seem beside the point. If science is really valuable, they argue, the market will step in. If scientists are really committed, they’ll make it work.

This is how a cultural narrative authorizes institutional harm. Not through dramatic confrontation, but through quiet permission to let things erode.

These narratives take many forms. Some frame public science funding as fiscally irresponsible in a time of national debt. Others argue that academic researchers should compete for private funding rather than receive government ‘handouts.’ Still others cast recent staff reductions at federal agencies as ordinary market corrections, no different from private sector downsizing. Each argument draws from the same cultural well: the presumption that government support for knowledge production is inherently suspect, that market discipline is the appropriate test of value, and that individuals who built careers in public service should simply adapt to new conditions. What these framings share is not malice, but a fundamental misunderstanding of what knowledge infrastructure requires and what happens when it’s treated as discretionary.

The Costs We Do Not Count

These narratives don’t just authorize cuts. They obscure what is actually being lost.

Forgoing sustained investment in basic biomedical science is not just a fiscal choice. It is a strategic one.

When public funding retreats, the U.S. does not simply save money. It cedes leadership to other countries such as China. It narrows the range of questions that can be asked. It reduces the likelihood that future breakthroughs will emerge here, or at all. The costs show up years or decades later, as missing treatments, delayed cures, and lost scientific capacity that cannot be quickly rebuilt.

Knowledge infrastructure, once dismantled, does not snap back into place.

An individualistic culture is poorly equipped to recognize this kind of loss because it unfolds slowly and diffusely. No single decision appears catastrophic. The damage accumulates quietly, until the system can no longer do what we assume it should.

Understanding this helps explain why contraction can be normalized even when the public broadly supports science. It is not only a political failure. It is a cultural one.

Markets tolerate churn. Knowledge systems do not. Once expertise is lost, it does not automatically return when funding improves. Once scientific capacity erodes, it cannot be instantly rebuilt. Once trust inside institutions is strained, it takes years to repair.

Why This Matters Now

The danger is not simply that science funding fluctuates. It is that a cultural narrative makes contraction feel morally acceptable, even prudent.

As I write this in February 2026, we’re watching this play out in real time. The H5N1 bird flu outbreak that began in dairy cattle has now spread to more than 70 human cases in the United States. The federal response capacity has been significantly weakened by staff losses, departures, and communication restrictions that limit how public health officials can share information with the press and public.

The research that might tell us whether this strain is developing human-to-human transmission requires the kind of long-term viral surveillance that looks like ‘waste’ on a spreadsheet. The expertise to interpret that data takes decades to develop and can be lost in days when experienced scientists are forced out.

This is the cost of treating public health infrastructure as negotiable overhead. Not someday. Right now.

When public institutions are judged primarily by how little they cost rather than by what they enable, rebuilding becomes politically harder, not easier. The goalposts move, and then they harden. What once would have been called underinvestment is rebranded as the new baseline.

Understanding this cultural backdrop does not require rejecting individualism outright. But it does require recognizing its limits.

Science has always been a collective endeavor. Treating it as though it should survive primarily through personal grit and private investment is not a neutral policy choice. It is a cultural one. And it shapes what we are willing to lose without calling it loss.

What a Democratic Vision of Science Would Ask

If this moment demands anything, it is not nostalgia for an imagined past era of abundance, nor faith that markets or heroism will quietly fill the gaps. It demands a clearer public conversation about what we believe science is for, how society benefits, and what it requires to function.

A more democratic vision of scientific investment would not ask only whether discoveries are exciting or budgets defensible. It would ask whether the structures that produce knowledge are durable, accountable, and worthy of collective trust. It would recognize that publicly funded science is not a luxury item to be trimmed during political discomfort, but a long-term civic commitment, like education or infrastructure, whose value compounds over time.

Such a vision would also resist the temptation to confuse endurance with success. It would not celebrate people for surviving instability while ignoring the costs of asking them to do so indefinitely. It would treat capacity, expertise, and institutional memory as assets to be stewarded, not excess to be squeezed.

A democratic vision of science would start by asking different questions:

Not ‘Can NIH make do with fewer staff?’ but ‘What knowledge will we never generate because those people are gone?’

Not ‘Did we avoid deep cuts this year?’ but ‘Can we recruit the next generation when faculty positions have been frozen for a decade?’

Not ‘Are scientists resilient enough to absorb this?’ but ‘Why are we asking them to?’

Not ‘Do Americans trust science?’ but ‘Does that trust translate into protecting the institutions that make science possible?’

It would recognize that some questions only get answered when researchers have the security to pursue decade-long projects with uncertain outcomes. That capacity cannot be rebuilt quickly once it’s lost. That ‘efficiency’ in knowledge production is often just a polite word for abandonment.

Most importantly, it would insist that when 77% of Americans say they trust scientists, that trust should translate into something more than warm feelings and viral social media posts praising healthcare workers. It should translate into the political will to maintain the institutions that make scientific work possible in the first place.

The question, then, is not simply whether Americans support science. The evidence suggests that they do. The question is whether we are willing to defend a version of science that cannot be reduced to quarterly returns, individual grit, or market investment alone.

What would it take to build public expectations that science deserves stability, not just admiration? What would it take to insist that collective and robust public investment, rather than quiet sacrifice, is the price of long-term progress?

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources

Kennedy, B., & Kikuchi, E. (2026, January 15). Americans’ confidence in scientists. Pew Research Center. https://www.pewresearch.org/science/2026/01/15/americans-confidence-in-scientists/

Cologna, V., Mede, N. G., Berger, S., et al. (2025). Trust in scientists and their role in society across 68 countries. Nature Human Behaviour, 9, 713–730. https://doi.org/10.1038/s41562-024-02090-5

Dzau, V. J., & Yamamoto, K. R. (2025). Government-funded health and biomedical research is irreplaceable. NAM Perspectives. National Academy of Medicine. https://doi.org/10.31478/202508b

Supran, G., Rahmstorf, S., & Oreskes, N. (2023). Assessing ExxonMobil’s global warming projections. Science, 379(6628). https://doi.org/10.1126/science.abk0063

Kupferschmidt, K. (2020, September 9). Why big pharma has abandoned antibiotics. Nature. https://www.nature.com/articles/d41586-020-02884-3

UNAIDS. (2025). Global HIV & AIDS statistics—Fact sheet. https://www.unaids.org/en/resources/fact-sheet

Rockey, S. (2014, November 20). The NIH funding crisis is really a biomedical research workforce crisis. Urban Institute. https://www.urban.org/urban-wire/nih-funding-crisis-really-biomedical-research-workforce-crisis

Nietzel, M. T. (2026, January 29). Hard times have come for the PhD degree. Forbes. https://www.forbes.com/sites/michaeltnietzel/2026/01/29/hard-times-have-come-for-the-phd-degree/

Hersher, M., & Mervis, J. (2026, January 27). U.S. government has lost more than 10,000 STEM Ph.D.s since Trump took office. Science. https://www.science.org/content/article/u-s-government-has-lost-more-10-000-stem-ph-d-s-trump-took-office

Herper, M., & Cueto, E. (2026, January 22). Unfilled vacancies have depleted NIH advisory councils, key players in grant approvals. STAT. https://www.statnews.com/2026/01/22/nih-advisory-council-vacancies-raise-questions-funding-politicization/

Centers for Disease Control and Prevention. (2026). H5 bird flu: Current situation. https://www.cdc.gov/bird-flu/situation-summary/index.html

In our All-Hands meetings before I left NIH, our leaders were always trying to think of ways to accommodate the new leadership, pitching ideas they thought would fit the “MAHA” agenda. It felt futile to me. They’d tell us that while things seemed uncertain, they were sure to get better—all of this while unannounced RIFs were stealing away hundreds of our scientific colleagues, grants were terminated, our external communications and travel were halted, and DOGE was surveilling us.

I know they meant well. They didn’t want to panic everyone or risk being fired. But I felt lied to. Not by malice, but by the refusal to acknowledge what was actually happening.

That’s when I realized we were being asked to pretend this was normal.

There is a particular moment many people in science recognize, even if they rarely name it out loud. It is the moment when something clearly feels wrong, but open resistance feels either futile or inappropriate. Funding is technically “stable.” Leadership sounds optimistic. Everyone is working harder than before to deliver less. The discomfort does not come from a single decision, but from the quiet realization that expectations have shifted, and that we are being asked to adjust our sense of what is acceptable.

This essay is about what institutions teach people to do in moments like this. Not through explicit instruction, but through repeated signals. When contraction is framed as resilience, when loss is reframed as efficiency, and when survival is treated as success, people learn how to absorb strain internally while performing normalcy externally. Over time, this reframing moves the goalposts. What once would have triggered alarm becomes a new baseline. What once felt untenable becomes merely difficult. Eventually, it becomes normal.

What makes the current moment in U.S. biomedical science especially destabilizing is not only the scale of change, but the way it is being narrated. New leadership emphasizes morale and “can-do” attitudes while downplaying the loss of staff and capacity. Advocacy organizations celebrate budget outcomes that do not repair structural damage. Meanwhile, the system itself is operating with fewer people, fewer awards, and fewer safeguards than it had two years ago. None of this requires anyone to lie. It relies instead on reframing, on selecting which facts are emphasized and which are quietly set aside.

The danger of this kind of normalization is not that it feels dishonest in the moment. It often feels responsible, even necessary. But each time the goalposts move without being named, expectations ratchet downward. People are trained to accept less support, less stability, and less institutional capacity as the price of commitment. Over time, the question is no longer whether science is being damaged, but whether we have been taught to stop recognizing the damage at all.

Optimism as a Management Tool

A recent interview in STAT with the newly appointed director of the National Cancer Institute, Anthony Letai, offers a clear illustration of how this reframing works in practice.

Asked about the coming year and the challenges facing the institute, Letai expressed confidence that NCI would fund more extramural grants than in the previous fiscal year. He also downplayed concerns about staff losses, emphasizing instead what he described as a strong “can-do attitude” among remaining staff. “I would be more worried if I didn’t see that can-do attitude,” he said. “I’ve been very impressed with the way people here have made do with fewer people and resources.”

On the surface, this kind of statement sounds supportive and appreciative. It acknowledges effort. It signals confidence. It reassures external audiences that the institution remains functional. But structurally, this kind of optimism does important work that is easy to miss if we focus only on tone.

First, it converts loss into virtue. Staff shortages and increased workloads are reframed as evidence of dedication, rather than as warning signs that capacity has been compromised. The ability to “make do” becomes a marker of success, even when the conditions that require it are the result of institutional harm.

Second, it shifts responsibility downward. When the emphasis is placed on attitude and adaptability, the burden of systemic failure is absorbed by individual workers. If things continue to function, it is because people rose to the occasion. If they do not, the implication is that the problem lies with effort or outlook, rather than with decisions that removed people, expertise, and time from the system.

Third, this framing closes the conversation. If making do is presented as success, then raising concerns about staffing levels or sustainability can be interpreted as pessimism or disloyalty. Structural critique becomes harder to voice precisely because the institution appears to be coping.

I don’t doubt Letai’s sincerity or his genuine appreciation for his staff. In 22 years at NIH, I said similar things myself during difficult periods. I know what it feels like to want to protect morale, to recognize genuine dedication, to avoid demoralizing people who are already stretched thin.

But I also know what it feels like to work on a team that’s “making do” with fewer people. The pride is real. So is the exhaustion. And the institutional choice to emphasize one while minimizing the other is how goalposts move. None of this requires bad intent. In fact, this kind of reframing is often motivated by a desire to protect morale and maintain legitimacy during difficult periods. But it is also how what once would have been named as unacceptable strain gets reclassified as admirable endurance.

When Victory Does Not Mean Repair

A similar dynamic is visible in how parts of the science advocacy community have responded to news about the NIH appropriations bill that recently passed the House and is awaiting a vote in the Senate. Several organizations issued statements celebrating the outcome as a meaningful victory for biomedical research, emphasizing that deep cuts were avoided and that Congress reaffirmed its support for science.

This response is understandable. Advocacy organizations operate in constrained political environments. Their role often involves protecting institutions from worse outcomes, maintaining access to decision-makers, and signaling stability to researchers who are already anxious. In moments of contraction, calling a fragile outcome a win can feel like an act of defense.

But this framing also moves the goalposts.

When a budget that does not keep pace with inflation is treated as progress, erosion of purchasing power becomes normalized. When the loss of staff, the reduction in awards due to multi-year funding schemes, and the hollowing out of institutional capacity are bracketed off as outside the scope of the victory, they quietly recede from the public narrative. Over time, the absence of further harm is reclassified as success.

Advocacy rhetoric shapes expectations. It influences what researchers, institutions, and the public come to see as reasonable. When stabilization without repair is celebrated, it becomes harder to argue for restoration. The system adapts to less, and then defends that adaptation as realism.

Administrative Silence as a Form of Normalization

The normalization of contraction is not driven only by messaging. It is also reinforced by what does not happen.

It’s not enough for Congress to appropriate dollars. The Department of Health and Human Services appears to be finding workarounds to curtail NIH science. Take, for instance, their stalling of approvals for NIH Advisory Councils.

By law, these councils must be in place to approve new funding. Yet according to a recent STAT investigation, since President Trump took office in January 2025, only one new member has been added to the NIH Advisory Councils. These are the key panels that make final recommendations on what research to fund at universities and medical centers.

As current members’ terms expired, a handful resigned, and HHS withdrew or denied other nominations, the councils have been depleted. The majority of the 25 councils are now operating with less than half their full complement of members. The vacancies have left many panels without the range of expertise needed to make informed funding decisions. Some appear in danger of running afoul of federal laws governing council makeup and operations.

I watched this happen at my own Center. Advisory Council vacancies went unfilled. No announcements. No explanations. Just a gradual erosion of the oversight structure that’s supposed to ensure accountability and scientific rigor in funding decisions.

This is not merely an administrative backlog. Advisory Councils are a critical safeguard. They provide scientific oversight, ensure accountability, and help maintain public trust in funding decisions. When councils are underpopulated or delayed, the system technically continues to function, but with diminished deliberation and transparency. Ultimately, if Advisory Councils lose all of their members, then by law grants for that Institute or Center cannot be awarded.

How will this situation be resolved? Will HHS speed up the review of the nominations? Will the terms of current members simply be extended over and over again? Or perhaps the Secretary of HHS, RFK Jr., will populate the Advisory Councils with his hand-picked cronies?

What is striking is how easily this kind of capacity loss is not being discussed amidst the backdrop of the celebrations about the NIH budget bill passing without cuts. There is no press conference announcing that NIH Advisory Council oversight is in danger of disappearing. There is no formal declaration that standards have changed. Instead, work proceeds more slowly, decisions are deferred, and the strain is absorbed by staff who are already stretched thin.

Here again, the goalposts move. What would once have been treated as an urgent governance failure becomes a background condition. The system adapts to operating with fewer checks, fewer voices, and less deliberation, and calls that adaptation resilience.

What Science Culture Teaches People to Do

Taken together, these examples point to a broader pattern. Science culture is very good at teaching people how to endure. It valorizes resilience, adaptability, and sacrifice. It treats overload as character-building and silence as professionalism.

During periods of expansion, these traits are rewarded. During periods of contraction, they become tools of normalization.

People learn to hold two realities at once: private recognition of structural harm, and public performance of optimism. They learn to interpret concern as weakness and endurance as virtue. Over time, this training changes not only what people say, but what they expect.

I felt this tension in every All-Hands meeting last year. I knew the pitch deck about “MAHA” alignment was theater. I knew we couldn’t accommodate our way out of what was happening. But I also understood why leaders were doing it. They were trying to buy time, protect people, maintain some operational space.

The problem is that this kind of accommodation teaches everyone to lower their expectations. Each time we nodded along and went back to work, we reinforced the idea that this level of dysfunction was manageable. That we could make it work. That it was normal.

What I Couldn’t Normalize

I realized I couldn’t pretend things would be okay at NIH, so I took the early retirement offer.

Not because I gave up on science. Not because I stopped believing in the mission. But because I understood that staying meant participating in the normalization of conditions I believed were unacceptable.

Every meeting where I didn’t speak up about canceled grants or depleted councils would be an endorsement of their depletion. Every time I emphasized staff resilience without naming the impossible workloads, I’d be moving the goalposts. Every budget cycle where I called flat funding “stable” would make it harder to argue that erosion was happening.

Leaving was not a solution. It was a recognition that I couldn’t be part of the performance anymore.

Refusing to Pretend This Is Normal

None of this is an argument for despair. It is an argument for clarity.

Normalization is powerful precisely because it feels reasonable. It asks people to be patient, flexible, and pragmatic. But when patience becomes acceptance and flexibility becomes acquiescence, real damage goes unnamed.

Anne Applebaum, writing about authoritarian regimes, notes that almost all of them try to undermine admired institutions in order to radically alter the way people think. Also, attacks on the civil service are a common tactic among autocrats—replacing neutral, merit-based experts with political loyalists to consolidate control, enforce obedience, and reduce bureaucratic resistance. The aim is to turn government institutions into tools for their personal agenda, enabling the purges of independent thinkers and rewarding cronyism. What I was watching in those All-Hands meetings wasn’t just budget management or organizational adaptation. It was the early stages of something more systematic: training the civil servants at NIH to accept a fundamentally different relationship between government and knowledge.

The cuts and the silencing of scientific staff aren’t just about money. They’re about power. About who gets to decide what counts as knowledge, who has the authority to produce it, and what conditions the public should accept as normal for the people who do that work.

Refusing to pretend that this is normal does not require outrage or certainty about solutions. It requires accuracy. It requires the willingness to say that loss is loss, even when people are doing their best to compensate for it.

It requires calling flat funding what it is: a cut. Calling staff losses what they are: a depletion of capacity. Calling the deferred filling of Advisory Council seats what it is: a weakening of oversight and accountability, and a possible backdoor to stopping grant awards from being issued.

Science depends not only on funding and infrastructure, but on honesty about conditions under which work is done. Holding that honesty, even when it is uncomfortable, is not pessimism. It is stewardship.

The consequence of failing to do this is not immediate or dramatic. It is gradual. We quietly accept a little less each time. Fewer staff. Fewer awards. Less oversight. Less transparency. Less capacity. More political control over science.

Until one day, we realize the entire federal biomedical research enterprise has been dismantled, not through a single catastrophic decision, but through a thousand small accommodations that we were trained to call normal.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

Sources:

STAT. Anthony Letai says NCI is more stable than perceived and is heading for a big year. https://www.statnews.com/2026/01/22/national-cancer-institute-director-anthony-letai-q-and-a/

Research!America Urges Quick Passage of FY26 Funding for Medical Research. https://www.researchamerica.org/press-releases-statements/researchamerica-calls-for-fy26-passage/

STAT. Unfilled vacancies have depleted NIH advisory councils, key players in grant approvals. https://www.statnews.com/2026/01/22/nih-advisory-council-vacancies-raise-questions-funding-politicization/

Anne Applebaum (2024). Autocracy, Inc.: The Dictators Who Want to Run the World. New York: Doubleday. https://www.penguinrandomhouse.com/books/725302/autocracy-inc-by-anne-applebaum/

Defund Science, Distort Culture, Mock Education: It’s not just about cuts to research. It’s about power. By Anne Applebaum. https://www.theatlantic.com/podcasts/2026/01/reality-reshaped/685289/

In 22 years at NIH, I learned to steward research portfolios through peer review and scientific judgment. If someone had told me in 2020 that political appointees would review all NIH grants, I would have quit. My job would have been rendered irrelevant.

That’s the reality facing NIH staff in 2026. Not just because of the funding bill Congress is advancing, but because of what they don’t yet know about how science will be funded going forward.

At moments like this, it’s tempting to read a funding bill as a verdict. Relief or disaster. Stability or collapse. The reality is more complicated. Congress appears poised to reject the most extreme proposals from the Trump Administration, and that matters. But the bill that looks like stability on paper leaves three profound uncertainties unresolved, each of which will shape what NIH can actually do in the coming year.

What the Administration Requested

To understand what Congress is preserving, it helps to be clear about what was originally proposed by the Trump Administration and the Office of Management and Budget.

The request included a roughly $20 billion cut to the NIH budget, a reduction of about 40 percent. It proposed eliminating four Institutes and Centers and consolidating the remaining 27 into just eight. It sought to dramatically expand a “multi-year funding” model so that nearly all grants would be awarded upfront for multiple years. And it aimed to severely restrict how NIH pays indirect costs to research institutions.

Taken together, these proposals would have fundamentally reshaped the scale, structure, and capacity of NIH. They would have reduced the number of grants awarded, narrowed the scope of supported science, and shifted financial risk onto institutions and investigators.

Some of these ideas were framed as efficiency measures. In practice, they would have represented a sharp contraction of public investment in biomedical research.

What Congress Is Likely to Do

Congress rejected the most extreme elements of that proposal. There will be no 40 percent budget cut. No institutes will be eliminated or forcibly consolidated. On paper, this looks like decisive preservation.

But preservation is not the same as repair. And what Congress is preserving comes with significant constraints.

The bill sets the NIH budget at approximately $48.7 billion, up only marginally from the $48.495 billion enacted in fiscal year 2025. Once inflation is accounted for, this is effectively a cut. The agency is being asked to do at least as much with less real purchasing power than it had the year before.

Congress also declined to give the Office of Management and Budget authority to apply multi-year funding to nearly all NIH grants. But it did allow the same approach used in fiscal year 2025, where roughly 30 to 40 percent of awards were funded upfront for multiple years.

Congress prevented NIH from changing how it pays indirect costs, preserving existing institutional support structures. That decision avoided an additional shock to universities and research hospitals.

So, the bill blocks collapse. It does not restore capacity. And it leaves three critical uncertainties unresolved.

First Uncertainty: How Many Research Teams Will Actually Be Funded?

Here’s what multi-year funding means in practice: Instead of funding five new R01 grants for their first year of work, NIH funds one grant for all five years. Four highly significant research projects—maybe a promising cancer treatment, a better approach to managing Type 2 diabetes—go unfunded. Not because peer review found them lacking, but because the math no longer works.

This isn’t hypothetical. In fiscal year 2025, when roughly 30 to 40 percent of NIH awards were funded upfront for multiple years, NIH made approximately 2,000 fewer awards than in the prior year, despite no reduction in the overall budget.

The appropriations bill allows that same approach to continue in fiscal year 2026, which means the number of funded investigator teams is likely to shrink further. Many strong proposals will go unfunded not because the science is weaker, but because fewer awards are available.

For researchers, this means heightened competition and longer odds. For NIH staff, it means even harder decisions about which grants to fund from an already impossibly competitive pool. Funding rates were already brutal in 2024. Now they will be lower still.

Some investigators may give up. Others may submit additional applications to increase their odds, creating even more review burden for scientific staff who are already overworked and understaffed.

The appropriations bill stabilizes the top line while permitting mechanisms that reduce the number of funded investigators, labs, and trainees. That’s not the same as stability for the research community.

Second Uncertainty: Can a Depleted Workforce Sustain This System?

Even if the budget were growing, the NIH that exists now is not the NIH of 2024.

During 2025, NIH lost roughly 4,400 employees out of a workforce of about 18,000. These losses came through formal reductions in force, probationary terminations, early retirements, buyouts, and separation offers.

Many of the people who left were not visible to the public, but they were essential to how the agency functioned. Program officers who understood portfolios in depth and could mentor investigators. Grants management specialists who could navigate complex policy rules. Scientific review officers who organized peer review panels. Analysts and support staff who kept systems running.

Those losses cannot be reversed by a single appropriations bill. Even with stable funding, fewer staff means higher workloads, longer timelines, and less capacity for thoughtful stewardship.

The remaining staff are already absorbing impossible workloads while navigating continuous procedural changes: new NOFO structures, centralized review, eliminated mechanisms, HHS approval requirements for every update and clarification. Now they face the prospect of making even more difficult funding decisions with even less support.

Funding can keep the lights on. It cannot instantly restore the human infrastructure that makes a complex institution work well. The appropriations bill provides no resources to rebuild capacity, no authorization to hire, no acknowledgment that institutional memory and expertise cannot be replaced overnight.

Third Uncertainty: Who Actually Decides Which Grants Get Funded?

This is the uncertainty that would have made me quit in 2020.

In August 2025, President Trump issued Executive Order 14042, “Improving Oversight of Federal Grantmaking.” The order significantly increases political control over NIH and other federal grants by requiring political appointee review for all funding decisions, prioritizing lower-cost institutions, and pausing new funding opportunities to implement new rules aligning grants with administration “priorities.”

The order potentially cuts funding for research involving diversity, equity, and inclusion initiatives. It increases scrutiny on certain research topics. And it fundamentally shifts power from scientific peer review toward political goals.

Here’s what no one knows yet: How will this actually work in practice?

Will every grant that peer review recommends for funding go through political appointee review? Will career scientific staff still make the final funding decisions, or will those decisions be overridden by appointees who may not have scientific expertise in the relevant areas? How long will the review process take? What criteria will political appointees use?

Current NIH staff haven’t been told whether the current approach of relying on peer review scores to select grants will continue, or if new political reviews will occur. The uncertainty itself is destabilizing.

For 22 years, I made funding recommendations based on scientific merit, feasibility, and alignment with public health priorities. Those recommendations went through peer review, program officer assessment, Council concurrence, and institute leadership review. The system wasn’t perfect, but it was grounded in scientific judgment.

If political appointees with different priorities are making final funding decisions, the role of program officers fundamentally changes. You’re no longer stewarding science. You’re processing applications through a review system whose outcomes you cannot predict or influence based on scientific merit alone.

That’s not the job I signed up for. And it’s not the job that attracted talented scientists to NIH for decades.

The Enforcement Problem

Here’s the deeper problem. Congress declined to give OMB authority to expand multi-year funding to nearly all grants. But what’s to stop OMB from directing NIH to issue a flurry of multi-year awards anyway?

With a Republican Congress, what oversight leverage exists? If the administration issues thousands of multi-year awards in defiance of congressional intent, are lawmakers going to force NIH to walk back those awards and re-issue them? The administrative chaos that would create—for institutions, investigators, and NIH staff—would be enormous.

This administration has demonstrated repeatedly that it doesn’t wait for authorization or follow established rules. Even when checked later by courts or Congress, ‘later’ can be months or years away. The chaos created in the interim may be the point. The appropriations bill cannot resolve this uncertainty because it cannot compel compliance.

What This Appropriations Bill Actually Signals

So, what should we take from this appropriations bill?

It signals that Congress is not ready to abandon NIH or radically dismantle it. It signals that the most extreme proposals were rejected. That matters.

It does not signal a return to stability. It does not reverse the effects of staff loss. It does not guarantee that strong science will be funded at prior rates. And it does not clarify how political control over grant decisions will actually function.

This is not a story of rescue or collapse. It is a story of partial preservation in a system facing three compounding uncertainties:

Funding uncertainty: Flat budgets and multi-year awards mean fewer research teams funded, even as costs rise.

Capacity uncertainty: Thousands of experienced staff are gone, and those remaining are absorbing impossible workloads with no clear path to rebuilding.

Authority uncertainty: Executive Order 14042 requires political appointee review of all grants, but no one knows yet how that will work in practice or whether scientific merit will remain the primary criterion.

For scientists and institutions, the practical implication is to be realistic without being fatalistic. Competition for grants will intensify. Timelines may lengthen. The criteria for funding may shift in ways that are not yet visible.

For staff inside NIH, the strain many are feeling is structural, not personal. The system is being asked to do the same work with less money, fewer people, and unclear authority over core decisions.

For those watching from outside, understand that rebuilding scientific capacity takes more than a line item in a bill. It requires people, institutional memory, stable governance, and clarity about who makes decisions and on what basis.

Looking Ahead

The appropriations bill keeps important pieces in place. It prevents the worst-case scenario. But it does not resolve the fundamental questions facing NIH in 2026.

How many research teams will actually be funded? Can a depleted workforce sustain the system? And who will ultimately decide which science moves forward?

Those uncertainties will play out over the coming year, in ways that funding numbers alone cannot predict or prevent. Understanding what this appropriations bill does—and what it cannot do—is essential for anyone trying to read this moment with clarity rather than illusion.

This is not a victory to celebrate. It’s a fragile holding pattern in a system under strain. Whether the biomedical scientific community can stay vibrant and employed under these conditions remains an open question.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they explore how science is shaped by funding, governance, and the people who steward it through periods of change.

I remember moments like this from my 22 years at NIH. A question might come up in a branch meeting about how to apply a particular statistical method to a multi-site trial, or whether a new AI approach was ready for clinical application. I could answer because I’d spent years building expertise across clinical research design, advanced methods, and practical implementation. And because I had a network across all 27 institutes I could leverage when needed to inform our deliberations.

Hundreds of Program Officers like me left the NIH in 2025. When those types of questions come up, there are fewer scientists who can answer them now.

But this isn’t just about scientific expertise walking out the door. It’s also the hundreds of communications staff who maintained relationships with patient advocacy groups, the grants specialists who knew how to solve complex award problems, the acting Directors who can’t provide stable leadership. This is a different kind of loss than budget cuts or canceled programs. It is harder to quantify, harder to explain, and much harder to reverse.

Over the past year, as the National Institutes of Health has moved through rapid restructuring and massive workforce reductions, this is the loss I think of most. It’s not confusion about whether specific innovations are ready, but the growing absence of the expertise and relationships that once made the system navigable.

The Knowledge That Rarely Lives in Documents

Large research institutions like NIH run on more than formal policy. They also rely on accumulated judgment.

Some of that judgment is technical. Much of it is interpretive. It includes knowing why a particular rule exists, when flexibility is appropriate, how to sequence decisions, and how to balance scientific opportunity against administrative risk. It includes understanding which statistical approaches work for which study designs, which AI applications are ready for clinical translation, and who across the agency has expertise in specific areas.

This knowledge is rarely written down in full. It is learned over years of watching programs evolve, seeing what fails quietly, and understanding where discretion matters most. It lives in conversations, in mentoring, and in the informal translation between policy intent and operational reality.

When people talk about institutional memory, they are not being nostalgic. They are describing a form of expertise that allows complex systems to function without becoming brittle.

The Scale of Loss

The NIH has lost roughly 4,400 employees out of a workforce of about 18,000. These departures include staff removed through formal reductions in force, probationary terminations, and approximately 2,000 people who left through early retirement, buyouts, or separation offers.

The losses include:

• Hundreds of communications staff

• Hundreds of program officers

• Hundreds of grants management specialists

• At least 11 to 13 Institute and Center directorships now vacant, with only acting Directors in place

These are the people whose work made the NIH the envy of the world: translating science for public audiences, stewarding research portfolios, processing complex awards, providing institutional leadership to champion discovery for the next treatment innovations and cures.

When restructuring happens at this scale and speed, two losses compound each other: the loss of people and the loss of institutional memory they carried.

What Leaves: Communications Expertise

Hundreds of communications staff were eliminated through the reductions in force. Previously, each of the 27 NIH Institutes and Centers had communications directors and staff with specialized knowledge in their specific disease areas, conditions, or scientific domains.

These staff did far more than write press releases. They:

• Coordinated newsletters for different audiences

• Created outreach materials translating complex science for patients, advocates, and the public

• Maintained and updated Institute-specific websites

• Issued press releases announcing new awards and key publications

• Facilitated webinars connecting NIH scientists with the research community

• Built and maintained relationships with scientific societies, advocacy groups, and patient organizations

• Responded to media and Congressional inquiries with scientific expertise

All of this specialized work, built around deep understanding of specific research areas, has been eliminated or centralized.

Now there is one centralized communications office for the entire NIH. The consequences are immediate and growing:

IC websites are no longer being updated. Information becomes outdated. Researchers looking for current program priorities find frozen content.

Far fewer communications and webinars are happening. The infrastructure for connecting NIH scientists with the research community has largely disappeared.

Outreach via social media is strictly controlled and centralized. The ability to communicate quickly and specifically about scientific advances in particular disease areas is gone.

Media inquiries are routed through HHS and the NIH Director’s office. Institutes no longer have capacity to respond to questions about research in their areas. Everything is filtered through political control.

Constituent relationships have been severed. Patient advocacy groups that worked with specific communications directors for years likely have no one to contact. The institutional knowledge of who cares about what research, which findings matter to which communities, and how to translate science for different audiences walked out the door.

Important findings don’t get communicated because there’s no one with the expertise to recognize their significance or the relationships to disseminate them effectively.

What Leaves: Program Officer Knowledge

Program officers steward research portfolios. They don’t just process applications. They understand scientific contexts, see connections across studies, mentor investigators, and help translate broad priorities into fundable research programs.

I carried knowledge about advanced statistical methods for clinical research, data science applications, AI approaches for clinical trials, and behavioral research. I had personal contacts across nearly all 27 Institutes that I could leverage to help my Center build collaborations, find relevant expertise, or coordinate across programs.

When I left, that knowledge left with me. I don’t know who answers those questions now, or whether anyone has the cross-cutting expertise and relationships to bridge those areas.

Hundreds of program officers are gone. The remaining staff have inherited impossible workloads. They are trying to steward portfolios that previously required multiple people, while simultaneously navigating the continuous changes described in earlier essays: new NOFO structures, centralized review, eliminated mechanisms, HHS approval requirements.

They are doing essential work under unsustainable conditions, without the institutional memory that once helped them navigate ambiguity.

What Leaves: Grants Management Expertise

Grants specialists process awards, manage budgets, ensure compliance, and solve the practical problems that arise when policy meets reality. They understand how rules work in practice, not just in theory.

Hundreds of grants management specialists are gone. Those who remain are doing work that has become significantly more complex and more fraught.

They are being asked to issue multi-year awards, which require substantially more upfront work and accounting. They were forced to terminate and reinstate thousands of awards following lawsuits challenging the administration’s actions. They may be ordered to terminate more awards in the future, regardless of scientific merit.

Their work has become highly precarious. They are navigating legal uncertainties, political pressures, and administrative complexity with far fewer people to share the load. The expertise about how to handle edge cases, resolve conflicts between policies, or work through implementation problems is stretched impossibly thin.

The Leadership Vacuum

At least 11 to 13 NIH Institute and Center directorships are now vacant, filled only by acting Directors.

Acting Directors are trying to hold down the fort and keep the trains running while staying out of the crosshairs of the administration. They cannot provide the kind of stable leadership, strategic vision, or institutional advocacy that permanent Directors could offer.

They cannot push back on unreasonable demands. They cannot take risks or make long-term commitments. They cannot protect their staff with the authority that comes from being a confirmed, permanent leader.

The remaining staff are absorbing impossible workloads without leadership that has their backs.

Why These Losses Compound

These losses do not happen in isolation. They create cascading failures.

Science cannot be communicated because hundreds of communications staff with specialized expertise are gone. Important findings go unannounced. Constituent relationships dissolve. Public understanding of NIH research erodes.

Research portfolios cannot be stewarded optimally because hundreds of program officers are gone and those remaining are managing multiple portfolios under constant procedural change. The nuanced guidance, mentorship, and strategic program building that once happened is less sustainable.

Awards cannot be processed efficiently because hundreds of grants specialists are gone, and those remaining are managing increased complexity (multi-year awards, potential terminations) under precarious conditions.

Institutes, offices, and centers cannot lead or adapt because acting Directors lack the authority, stability, or protection to advocate for their institutes or their staff.

Each loss makes the others harder to absorb. The system is not collapsing, but it is operating at a significantly diminished capacity.

Why These Losses Are Hard to See at First

Loss of institutional knowledge rarely shows up immediately in metrics.

Applications are still reviewed. Awards are still made. Reports are still filed. From the outside, the system appears to be working.

The effects emerge gradually. Programs become less coherent. Decisions feel more risk averse. Mistakes that were once avoided begin to recur. People spend more time interpreting rules and less time thinking strategically. Opportunities are missed because no one recognizes them. Questions go unanswered because the people who held the answers are gone.

By the time these patterns become visible, rebuilding the lost expertise is no longer straightforward. The people who carried it have moved on. The conditions that allowed it to develop no longer exist. The relationships that made informal problem-solving possible have been severed.

Rebuilding Without Memory Is Harder Than Starting Over

There is a common assumption that if funding returns or policies stabilize, institutions will simply revert to what they were before. That assumption underestimates the role of memory and relationships.

Institutions that lose experienced staff do not reset to an earlier version of themselves. They become something new, often less capable of learning from past experience.

Rebuilding scientific infrastructure requires more than appropriations. It requires people who understand how parts fit together, why certain safeguards exist, and how to balance competing values under uncertainty. It requires networks of relationships built over years. It requires accumulated judgment about what works, what fails, and why.

Once that knowledge is gone, it cannot be replaced quickly. It must be rebuilt through time, trust, and repeated practice. The communications relationships with patient advocacy groups, the cross-Institute networks that enabled collaboration, the grants management expertise about how to handle complex situations—none of this can be immediately restored with a brand-new hire and a training manual. (Moreover, the agency is operating under a hiring freeze at present!)

Why Naming This Loss Matters

Naming this kind of loss is not an argument against all change. Institutions must evolve, and some restructuring may be necessary.

But change without attention to what is being lost risks weakening the very capacity the system will need in the future.

For those still inside NIH, recognizing this dynamic can help explain why the work feels harder. The complexity hasn’t disappeared. The people who helped navigate it have.

For those who have left, it affirms that what we carried mattered, even if it was never formally acknowledged. The statistical expertise I built over decades. The relationships I maintained across institutes. The judgment about when to push for innovation and when to proceed cautiously. That knowledge had value, and its absence creates gaps that are not easily filled.

And for anyone thinking about how science might be rebuilt in a more stable future, this is a lesson worth holding onto. Infrastructure is not only policies and platforms. It is people, judgment, memory, and relationships.

When experienced people leave, those are often the first things to go. And they are the hardest things to rebuild. Will the country’s leaders find a way to invite back the deep knowledge and experience of scientists like me who left if and when the time comes to rebuild?

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they examine how science is shaped by the systems, structures, and people that sustain it.

A program officer I know is anticipating total confusion. She manages clinical trials research, and she can already see what’s coming when the new NIH Notice of Funding Opportunity (NOFO) policies take effect. Investigators will try to propose trials tailored to specific disease areas or intervention phases, but the NOFOs will only accommodate a centralized, one-size-fits-all approach.

Many NIH institutes and centers have carefully developed clinical trial stage frameworks. They guide investigators to consider the appropriate level of inquiry: Is this a feasibility trial testing whether a brand-new intervention can actually be delivered in a specific setting or with a particular patient population? An efficacy trial establishing that it does work under controlled conditions? Or an effectiveness or pragmatic trial determining whether it works in real-world practice?

These trial types are fundamentally different in size, purpose, and cost. Lumping them together in generic parent NOFOs will create confusion for everyone: program officers trying to steward research programs, scientific review officers organizing panels, reviewers evaluating applications, and investigators trying to figure out where their work fits.

She’s bracing for the questions she may not be able to answer clearly. Not because she doesn’t understand the science, but because the new system eliminates the grant mechanisms that used to provide clarity. She hesitates further because she does not want to provide favoritism for funding applications that are specific to the institute where she works, the subject area in which she maintains expertise, and she is in the dark as to what guidance to provide regarding NOFO responsiveness.

There is a particular kind of exhaustion that comes not from one big disruption, but from many small ones arriving all at once. Each change on its own might seem manageable, even reasonable. Together, they create a sense that the ground is constantly shifting, and that keeping up requires more effort than most people can sustainably give.

That is the feeling many people inside the biomedical research enterprise are experiencing right now. Not confusion about a single new rule, but fatigue from the sheer volume of adjustments required to keep doing ordinary work.

In 22 years at NIH, I watched the agency absorb many changes. But I never saw this many arrive simultaneously with this little preparation.

Why Simplification Can Feel Harder, Not Easier

In periods of reform, simplification is usually presented as an unambiguous good. Fewer forms. Fewer pathways. Fewer exceptions. Fewer prior approvals required. In theory, simplification reduces friction and frees people to focus on substance rather than process.

In practice, simplification often shifts complexity rather than eliminating it.

The NIH is implementing sweeping changes under the banner of “administrative burden reduction.” The number of Notices of Funding Opportunity will be cut by 50% compared to last year. Different types of funding announcements (RFAs, PARs, PASs) will be standardized into uniform NOFOs. Modular budgets will be eliminated. Data management plans will be simplified to three yes/no questions. Institute-specific guidance pages will be replaced with centrally managed websites. Individual program officer contacts will be replaced with group email addresses.

On paper, this reduces fragmentation. In practice, it shifts responsibility outward. Investigators and institutions must now interpret how general guidance applies to highly specific scientific, administrative, and clinical contexts.

Take the clinical trials problem my colleague is facing. For years, institutes developed sophisticated frameworks for different trial phases. A feasibility trial for a behavioral intervention to test whether the intervention can be delivered in community health centers might be done with an R34 with direct costs requests of up to $450,000 over the entire project period and only $225,000 annually. A multi-site efficacy trial of the same intervention might need $350,000 in direct costs per year with an R01 over 5 years. A pragmatic trial testing real-world implementation might need a phased UG3/UH3 mechanism with millions over six years and involve many more sites.

These are not interchangeable. They require different expertise to review, different infrastructure to support, and different timelines to complete. Collapsing them into generic parent NOFOs doesn’t eliminate that complexity. It just makes it harder for everyone to navigate.

Or consider what happened with data management plans. Over the past couple of years at NIH, I was one of the program officers helping to shape and implement the Data Management and Sharing Plans policy. It took years to develop. We worked with institutions, investigators, and data repositories to create clear expectations. We trained program staff. We answered hundreds of questions as investigators learned to write these plans.

Now it’s being dismantled and replaced with three yes/no questions.

This is whiplash. For the program staff who spent years crafting policy. For the investigators who finally understood the requirements. For the institutions that built infrastructure to support data sharing.

The complexity of data sharing hasn’t disappeared. Someone still has to figure out repositories, timelines, access controls, and privacy protections. The difference is that now those decisions happen with less guidance and less institutional support.

This is not a failure of simplification. It is a predictable consequence. The question is not whether simplification is desirable, but whether the system has capacity to absorb the interpretive burden it creates.

What Centralization Changes

Alongside simplification, the changes implement wholesale centralization of information and authority.

Institute-specific guidance pages are being eliminated. Individual program officer contacts are being replaced with group email addresses.

This sounds administrative. It’s not.

When NOFOs were published, investigators could identify the program officers who wrote them. Names and email addresses were listed. An investigator with questions about scientific scope, budget justification, or whether their preliminary data was sufficient could reach a specific person who understood that research area. Those connections are being severed. Investigators will now have only group email addresses and no names.

What happens when an investigator emails a group address with a complex question? Someone will respond, presumably. But will it be someone with expertise in that specific research area? Will it be someone who knows the history of that program and why certain decisions were made? Will the response be authoritative, or will it need to be escalated through approval chains?

For decades, the NIH system worked in part because it preserved direct connections between scientific experts. Program officers weren’t just administrators. They were scientists who could provide nuanced guidance because they understood both the research and the funding landscape.

Group emails replace expertise with scalability. The system may handle higher volumes of generic questions, but it loses capacity for the specialized judgment that complex science requires.

The HHS Approval Constraint

What makes these changes particularly rigid is a constraint that rarely appears in public communications: every single change now requires HHS approval.

Not just major policy shifts. Every link update. Every contact change. Every piece of information NIH wants to post or clarify must be cleared by the Department of Health and Human Services.

This transforms what might be a challenging reorganization into something more restrictive. NIH cannot adapt quickly to emerging needs, cannot clarify ambiguous guidance in real time, cannot respond to investigator questions with updated information unless those updates clear political review.

Many institute websites are already outdated because there is no clear pathway to update them. Program officers who once could post clarifications now must submit requests up multiple approval layers and wait.

When I was still at NIH, I could update program information, post new resources, or clarify requirements when investigators raised questions. That responsiveness is no longer possible. The approval bottleneck affects everything.

What Gets Lost: The 200-Character Problem

Perhaps the clearest example of what’s being eliminated is how institutes can now communicate scientific priorities.

For IC-issued NOFOs, each participating institute will only be allowed to provide up to five bullets indicating areas that may be prioritized for funding. Each bullet is limited to 200 characters.

Two hundred characters is the length of a text message. Try describing a scientific priority in 200 characters. The nuance, the context, the rationale for why this area matters now—none of it fits.

And here’s the critical detail: these priorities “will have no referral/review impact.”

So, institutes can technically state what they care about, but it won’t affect how applications are reviewed or where they’re assigned. The priorities are decorative.

Program officers who spent careers understanding gaps in knowledge, emerging opportunities, and strategic directions can now communicate that expertise in five sentences of 200 characters each. And it will be ignored during review.

This is the systematic removal of scientific judgment from funding decisions.

Why Volume Matters More Than Any Single Change

What makes this moment particularly destabilizing is not any individual reform, but their simultaneity.

NOFOs cut by 50%. Modular budgets eliminated. Data plans restructured from detailed policy to three yes/no questions. Guidance pages dismantled. Individual contacts replaced. Clinical trial policies centralized. Institute-specific flexibility removed. Every clarification routed through HHS approval.

Any one of these could be absorbed. Together, they leave little time for adjustment before the next shift arrives.

When changes arrive at this pace, training lags behind implementation. Institutional memory becomes less useful because past experience no longer maps to current practice. People spend more energy figuring out how to operate than doing the work itself.

This is institutional overload. Not chaos, but continuous low-level strain.

What Science Loses When Systems Strain

Science is particularly vulnerable to this condition because it depends on long timelines, tacit knowledge, and coordination across many actors.

Building coherent research programs becomes harder. Program officers once designed sequences of funding opportunities that moved fields forward progressively. Early feasibility studies, followed by efficacy trials, followed by implementation research. That coordinated approach required different NOFOs for different phases. It’s no longer allowed.

Specialized expertise becomes harder to access. When review is centralized and investigators can only reach group email addresses, the match between scientific questions and appropriate funding mechanisms becomes more random. Good science may not get funded, not because it lacks merit, but because it couldn’t find the right pathway.

Rapid response becomes impossible. When every clarification requires HHS approval, the system cannot adapt to emerging scientific opportunities or public health needs.

Mentorship relationships dissolve. Early-career investigators learn to navigate the funding system through relationships with program officers who explain unwritten norms, provide feedback on ideas, and help them understand when they’re ready to compete. Group emails don’t mentor.

The clinical trials confusion my colleague is anticipating? That’s not a temporary adjustment period. It’s a permanent loss of capacity to organize research around scientific logic rather than administrative convenience.

What Periods of Overload Reveal

Periods like this surface truths that are easier to ignore during stability.

They reveal how much depends on invisible labor. The program officers who translate broad policy into actionable guidance. The scientific review officers who organize specialized panels. The administrators who keep projects moving despite uncertainty.

They expose persistent tensions. Between efficiency and stewardship. Between uniformity and context. Between accountability and trust. Between scientific judgment and political control.

These tensions don’t indicate failure. They indicate a system being restructured faster than it can adapt, under constraints that limit its capacity to self-correct.

What Happens Next

My colleague managing clinical trials research doesn’t know how the new system will actually work in practice. The policies are clear on paper. The implementation details are not.

Will investigators figure out which parent NOFO to use? Will review panels understand the distinctions between trial types without specialized guidance? Will program officers be able to steward coherent research programs when they can’t design targeted funding mechanisms?

These aren’t rhetorical questions. They’re genuine uncertainties that will play out over the next year as investigators try to navigate the new landscape.

What I know from 22 years inside the system is this: complexity doesn’t disappear when you simplify structures. It relocates. Sometimes that relocation is productive, distributing problem-solving across many actors. Sometimes it’s destructive, creating confusion and delay that compound over time.

The test will be whether the system can absorb this much change this quickly, or whether the overload becomes chronic—wearing down the people trying to make it work until something essential breaks.

This essay is part of an ongoing series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise. Each piece stands alone, but together they explore how science is shaped not only by ideas and funding, but by the structures that support or constrain them.

A former NIH colleague texted me in despair. A research program she helped build and manages was being scrutinized and at risk of cancellation. She would have to spend weeks negotiating with investigators to change titles and specific aims of already-awarded studies. Every mention of race had to go. Every reference to structural barriers had to be scrubbed. The science hadn’t changed. The intervention hadn’t changed. But the words that accurately described the work were now forbidden.

She probably spent countless hours “sanitizing” her grant portfolio. Not for scientific reasons. For political ones.

For much of a scientific career, the system provides signals about what matters. Funding suggests value. Institutional affiliation suggests legitimacy. Advancement suggests progress. When those signals are steady, they carry much of the burden of judgment for us.

But what happens when those signals become unreliable or actively misleading? When funding decisions reflect political calculation rather than scientific merit? When the system can no longer tell you what counts?

That responsibility shifts back onto individuals, often abruptly and without guidance. This essay is about how people are navigating that shift right now.

When Systems Carry Judgment

In times of expansion, research systems quietly perform a crucial function. They translate collective priorities into individual validation. Grants, titles, and institutional roles become proxies for worth, impact, and direction. This arrangement is imperfect, but it works well enough that most scientists can focus on their work without constantly reassessing their legitimacy or purpose.

When systems shrink or become politically hostile, those proxies break down. Good work goes unfunded. Programs disappear despite demonstrated value. People who have spent years stewarding complex efforts find themselves without a place to stand.

The result is not just professional disruption, but moral disorientation. This is why contraction feels so destabilizing. It’s not only the loss of resources. It’s the loss of shared signals about what counts.

Three Ways People Are Living This

The loss of reliable signals looks different depending on where you stand. Here are three versions I’m watching unfold.

The Program Officer Still Inside

My former colleagues still at NIH are discovering that the job they trained for is being devalued.

Program Officers used to design Notices of Funding Opportunities (NOFOs) to build particular areas of health science. They’d work with Scientific Review Officers to structure specialized review panels to ensure peer reviewers understood the nuanced science in their portfolios. They’d work with investigators over multiple phases of research, supporting teams through feasibility studies, efficacy trials, and implementation research with coordinated funding announcements.

Now, NOFOs are controlled by political operatives. Existing funding announcements have been canceled without scientific rationale. There are arbitrary limits on the total number of NOFOs the NIH can issue, regardless of scientific need. All peer review must be done centrally, eliminating the specialized study sections that Scientific Review Officers carefully built to evaluate specialized research grant applications.

The phased approach to clinical trials (supporting teams through early feasibility testing before advancing them to efficacy and effectiveness or pragmatic trials) is being threatened. NIH no longer allows multiple linked NOFOs that reflect the different phases of clinical trials research simply because leadership imposed an arbitrarily set limit to the number of NOFOs published annually. NIH also will no longer allow specialized reviews tailored to phased programs as all review is now centralized. This may end up seriously impeding progress for the health intervention research pipeline across the institutes and centers as a one-size-fits-all model often won’t work for different conditions, different diseases, and diverse kinds of interventional approaches.

And there are clear signals that peer review may become less important for funding decisions going forward. Political appointees may have the final say instead of career scientists. Certain science, particularly research on health inequities and marginalized populations, may be prevented from moving forward regardless of scientific merit.

Program Officers may need to negotiate with investigators to change already-awarded studies. To remove language about race. To scrub references to structural barriers. To make the work politically palatable rather than scientifically accurate.

The core function of the job (scientific stewardship) is under assault. What may remain is a bureaucratic role stripped of judgment, expertise, and autonomy.

Yet some Program Officers are staying because someone needs to carry institutional memory forward. Because leaving means no one will know why systems were designed the way they were, or what was learned along the way. Some may still be holding out hope that the dismantling of the system will be halted by the courts or by congress so that the work can return to some sense of “normalcy.” Either way, they’re redefining success from growth to damage mitigation. From building to holding the line.

The Early-Career Researcher Whose Future Disappeared

In early 2025, the NIH MOSAIC program (Maximizing Opportunities for Scientific and Academic Independent Careers) was terminated. The program supported early-career scientists from diverse backgrounds, providing not just funding but mentorship, training, and a pathway to independence.

The cancellation came as part of broader federal efforts to eliminate so-called Diversity, Equity, and Inclusion initiatives. The administration viewed MOSAIC as conflicting with its directive to end DEI programs, even though these awards were all earned through rigorous scientific merit review.

For the awardees, the impact was immediate and devastating. Grants were cut mid-stream. Research halted. Lab setups were disrupted. Career trajectories that had taken years to build collapsed overnight.

Other programs followed: U-RISE, G-RISE, and MARC were also cut. The NIH later extended eligibility for K99/R00 grants due to broader application delays, offering some relief. But for MOSAIC recipients, there was no comparable alternative. They’d been selected for a program designed to support scientists from disadvantaged backgrounds (a definition that went beyond race and ethnicity) and now that pathway was simply gone.

These researchers are facing an impossible calculation. They trained for a system that valued their perspectives and their science. That system is in tatters. Some are scrambling for new funding. Some may decide to leave science altogether. Some may try to wait it out and hope conditions will change before their expertise atrophies or their resources run out.

What do you do when the external signal that validated your career is not just withdrawn, but actively repudiated? When the message becomes: your science is not wanted here? You’re left to figure out your own compass, often with no institutional support and no clear path forward.

The Departed Steward: My Own Reckoning

Since leaving NIH in April 2025, I’ve felt a loss of identity that I didn’t anticipate.

For 22 years, I was a health scientist. I built programs, championed methodological innovations, mentored colleagues, coordinated multi-institute initiatives. My work was validated by the institution, by the grants I managed, by the influence I had on the direction of research portfolios.

Now, when people ask what I do for a living, I fumble.

I’ve applied for numerous research positions in the private sector. I haven’t landed a job yet that matches my skill set, vision, and expertise. I’m beginning to wonder if I need to put my scientific career on hiatus.

I’m still contributing to scientific advocacy efforts. I’m helping teams draft papers and books. But I don’t have any key scientific initiatives that I’m responsible for managing anymore. The work that defined me for over two decades (stewardship, program building, coordination) doesn’t translate easily outside the institution that housed it.

What surprises me most is how much of my identity was tied to external signals I didn’t realize I was relying on. The NIH email address. The ability to convene experts. The authority to shape funding priorities. The sense that my judgment mattered because the institution validated it.

Without those signals, I’m discovering what’s actually portable and what was always system-dependent.

Some things remain. The methodological knowledge I built over decades. The ability to see connections across disciplines. The judgment about what makes research rigorous and trustworthy. The commitment to serving public health, even if I’m not sure how to do that anymore.

But other things have evaporated. The daily sense of purpose. The feeling that my expertise is being put to use in ways that matter. The identity of “health scientist” that once felt so solid.

I’m learning, slowly and uncomfortably, that orientation matters more than institutional affiliation. That knowing what I stand for is steadier ground than any job title. But that learning is harder than I expected, and lonelier.

What Actually Endures

When external validation becomes scarce or distorted, some forms of value endure more reliably than others. Watching these three stories unfold (the Program Officer trying to preserve what’s possible, the early-career researcher whose future vanished, and my own uncertain path) a few patterns become clear.

Methods that produce usable truth still matter. Not just results that can be published, but approaches that hold up under uncertainty, political pressure, and real-world complexity. The Program Officer sanitizing grant titles can’t change what the intervention actually does. The research design is still sound. The data will still be trustworthy. Work that can be trusted across contexts retains value even when the system around it falters.

People who carry knowledge across cycles matter. Institutions forget faster than individuals do. The Program Officers still inside, the MOSAIC researchers who stay in science despite everything, the departed staff who remember why systems were built the way they were: these people become especially important when continuity is threatened. They’re the bridges.

Capacity-building work matters, even when it’s undervalued. Governance design, data infrastructure, coordination mechanisms, mentoring are often the first casualties of contraction. Yet they’re precisely the elements that make rebuilding possible later. The Program Officers who stay aren’t just processing paperwork. They’re maintaining architecture that will matter if conditions ever change.

Finally, integrity under pressure matters. When incentives misalign, shortcuts become tempting and silence can feel safer than candor. Choosing to uphold standards, even when doing so carries personal cost, is a form of stewardship that outlasts any single funding cycle. The colleague who texted me could have just changed the grant titles without question. Instead, she negotiated with the investigators to preserve as much scientific accuracy as possible while meeting political demands. That kind of judgment (knowing where to hold the line and where to adapt) is what endures.

Staying, Leaving, and False Moral Binaries

Periods like this sharpen difficult choices. Some people remain inside federal science institutions, working under increasing constraint. Others leave, whether by choice or necessity, taking their expertise elsewhere.

Neither path carries inherent moral superiority. Staying is not automatically complicity, and leaving is not abandonment. The real question is whether one’s work, wherever it’s done, continues to be guided by judgment, care, and responsibility to evidence.

I left because I couldn’t participate in what was being demanded of the institution. Many of my program colleagues stayed because they believe preservation work is critical. The MOSAIC researchers who leave science aren’t failures. They’re responding rationally to a system that told them they weren’t wanted. The ones who stay and try to wait it out aren’t naive. They’re betting on resilience.

What matters is not proximity to the system at any given moment, but whether knowledge, standards, and ethical commitments are preserved across transitions.

Why Orientation Matters More Than Optimism

It’s natural to look for reassurance during periods of instability. History suggests that rebuilding phases often follow contraction, but they rarely restore systems exactly as they were. Institutions return altered, and not all losses are recovered.

In that context, optimism is less useful than orientation. When the science system is unstable, it becomes harder to rely on external markers of success. It becomes easier to lose sight of why the work mattered in the first place.

Holding the line on meaning is not a sentimental act. It’s a practical one.

My program officer colleagues who spent weeks sanitizing grant titles are still stewards, even when the work feels compromised. The MOSAIC researchers who are scrambling for new funding are all still scientists, even without institutional validation. I’m still a health scientist, even when I fumble to explain what that means now.

I don’t know if the Program Officers still inside will see their programs restored. I don’t know if the MOSAIC researchers will find a place in science or have to rebuild their careers elsewhere. I don’t know if I’ll work as a health scientist again.

But I still know what I believe good science requires: rigorous methods, ethical stewardship, commitment to evidence, and service to public health. Those commitments don’t depend on institutional validation. They’re portable, even when everything else is in flux.

Careful methods, institutional memory, capacity building, and integrity don’t disappear when funding contracts. They remain, carried by people, waiting for conditions in which they can be expressed more fully again.

This essay is part of a series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise.

Five hundred people. One virtual meeting. A handful of HR staff trying to explain retirement paperwork to a room full of departing colleagues—while some of those same HR staff were being terminated themselves. They broke down crying together on screen. That was my last day at the National Institutes of Health (NIH): April 18, 2025.

Many people in the scientific community are grieving right now. Grants have been delayed or canceled. Programs are being scaled back. Careers built over decades suddenly feel fragile or dispensable.

If you work in or around federally supported science—as a civil servant, a grantee, a trainee, or a contractor—this moment may feel less like a policy shift and more like a personal reckoning.

This essay is not about assigning fault or predicting collapse. It’s about understanding where we are in a longer historical cycle, and what it means to hold the line during periods when public support for science contracts rather than expands.

Who This Is Landing On

I want to be explicit here. This moment is landing hardest on federal science workers: program officers, scientific review officers, intramural researchers, analysts, and support staff whose work has quietly sustained the research enterprise for decades.

I think about my former colleagues still inside, watching programs they built contract in real time. The program officer who designed a multi-year initiative that will now never launch. The grants specialist who drew up complex awards, only to be told later they must be rescinded. The scientific review officer now overwhelmed by full consolidation of review in the Center for Scientific Review, but with less staff to handle the workload.

It’s also landing on investigators whose labs and trainees depend on federal grants, who are now facing uncertainty through no failure of their own. The researcher whose renewal is stalled. The early-career investigator whose first award got canceled. The senior scientist who is forced to lay off their trusted research staff as grant awards sunset and new ones become increasingly more difficult to obtain.

I left NIH in April 2025 through the Voluntary Early Retirement Authority. I chose to walk away from 22 years of work I loved rather than participate in what was being demanded of the agency.

The sense of abandonment many are feeling is real, and it deserves to be named.

Contraction Is Part of the History, Even If It Never Feels That Way

When you’re inside a contraction, it doesn’t feel like history. It feels like failure: personal failure, institutional failure, maybe even moral failure.

But periods of contraction are not new in the history of science. What is new is how rarely we talk about them openly while they’re happening.

American science has expanded and contracted over time, often in response to political realignment, war, economic pressure, or shifting public trust.

Before World War II, the United States was not yet the dominant scientific superpower it would become after 1945. Federal investment in civilian research was relatively limited and concentrated in mission areas such as agriculture and resource surveying, while universities and laboratories relied heavily on philanthropic foundations and industry, which in turn influenced which scientific questions were pursued and which were neglected. ^{1, 2, 3}

The massive postwar expansion that many of us implicitly treat as “normal” was historically unusual. In the decades after World War II, rapidly growing federal support for basic research, especially in the context of the Cold War and the Space Race, created an era of “Big Science,” but even that era saw periods of constraint. In the 1970s, as growth in federal research budgets slowed, oversight increased, and economic turbulence hit, funding became tighter in several fields, and scientific careers—particularly in areas such as high‑energy physics and some branches of engineering—became less secure, prompting some highly trained researchers to leave academic science altogether. ^{4, 5, 6, 7}

President Bill Clinton, along with bipartisan congressional support, initiated and oversaw a major increase in the NIH budget, resulting in its doubling from around $14 billion in 1998 to over $27 billion by 2003, a period marked by significant annual funding hikes, often exceeding 14-17%, to accelerate biomedical research and tackle diseases like cancer and map the human genome. ^8,9,10

What makes the current contraction feel especially destabilizing is that it’s not driven by a broad loss of public support for science. Polling consistently shows strong majorities favor federal investment in health research. The dismantling we’re witnessing is instead the result of a political moment in which a minority viewpoint holds disproportionate power.

That distinction matters. When contraction reflects shifting public priorities, rebuilding can follow familiar paths. When it reflects ideological hostility to science itself, the damage isn’t easily reversed, even if political control later changes hands.

Why Contraction Feels So Personal

Even when contraction is systemic, it lands personally.

Federal science workers often experience these moments as a loss of identity as much as a loss of employment. For investigators, a paused or unfunded grant isn’t just a budget problem. It’s a threat to trainees, collaborators, and years of accumulated expertise. For program staff, watching carefully designed programs stall or unwind can feel like professional grief.

The system rarely acknowledges this emotional reality. Science culture valorizes resilience and adaptability without recognizing the real costs of constant instability. People are expected to pivot, retool, or “find another funding source,” as though that were simply a technical adjustment rather than a profound disruption.

In early 2025, the NIH experienced a freeze. No grant awards allowed, no meetings, no travel, no outside communications. Then came the February 14th illegal firing of over a thousand colleagues with no prior warning to staff or supervisors, followed by immediate lockout from systems.

By March, those of us still inside faced mandatory daily office presence, increased campus police searches, constant DOGE surveillance, and petty rules. Immediate termination was threatened for not displaying ID badges or powering down laptops, even when using the restroom. We were required to submit weekly activity reports to DOGE under threat of dismissal. Even more shockingly, some of my NIH colleagues were cruelly named on a “watchlist” posted online by the so-called “American Accountability Foundation,” which wanted to remove them from their jobs for allegedly “promoting liberal ideologies.” The objective: terrorizing and pushing out dedicated federal employee scientists.

The April 1st Reduction in Force brought even greater chaos. Staff were crying in the halls. Critical personnel across communications, IT, budget, and leadership were dismissed with no transition. The HR staff who were fired were tasked with off-boarding themselves and thousands of others.

All while we were expected to maintain scientific integrity and serve the public mission.

Naming this matters. Not because it solves the problem, but because silence compounds it.

What Contraction Exposes

Periods of contraction reveal truths that are easier to ignore during expansion.

They expose how much of the system depends on invisible labor. When program officers disappear, when institutional knowledge walks out the door, when support staff are cut, the gaps don’t show up immediately. They show up weeks later when critical coordination doesn’t happen, when protocols aren’t enforced, when the people who knew how to process payments are simply gone.

They highlight which kinds of work are considered essential and which are treated as expendable. Stewardship work (building consensus across institutes, mentoring colleagues, establishing governance structures) is almost always considered expendable, right up until its absence creates chaos.

They also clarify the difference between rhetoric about “valuing science” and the structures that actually sustain it. On April 1st, the day thousands of HHS employees were fired, the new NIH Director sent his first email to staff, thanking departing colleagues and claiming their firing was “in no way a reflection of the quality of their work.”

But it was. They were fired because their work was invisible until it was gone. Because stewardship doesn’t show up on efficiency metrics. Because expertise takes years to build and seconds to dismiss.

The Bridge Forward Is Not Just About Money

It’s natural to hope for a rebuilding phase, and history suggests that such phases do come. But rebuilding is never a simple return to what was.

The most consequential question isn’t whether funding will eventually increase again. It’s how the next system is designed and what it learns from the period of contraction that preceded it.

What do we actually want public science to do? What kinds of work requires stable, long-term support? What roles are essential to stewarding knowledge, not just producing it?

The dismantling is still ongoing, which makes it hard to imagine rebuilding yet. But one thing seems clear: we will need to find ways to leverage and connect the popular support for health research that polling consistently shows exists. We may need more explicit ways to directly engage the public in the work and connect the benefits of science with the American people more regularly. Support for public science cannot depend solely on political winds when those winds can shift so dramatically against public sentiment.

Moments like this create space, however uncomfortable, to rethink assumptions that were previously locked in by abundance. During expansion, it’s easy to layer on new programs without questioning foundational structures. During contraction, those questions become unavoidable.

For those still inside the system, this may mean holding two realities at once: continuing to do careful, ethical work under constrained conditions, while also imagining different structures that could support science more sustainably.

For those pushed outside it, and I’m one of them now, the loss is real. So is the expertise that risks being squandered. People who spent decades building methodological innovations, mentoring colleagues, establishing research programs don’t simply carry all that institutional memory with them when they leave. Much of it just disappears.

Recognizing that loss is part of any honest rebuilding.

Holding the Line on Meaning

I can’t offer certainty about timelines or outcomes. What I can offer is this: contractions don’t erase the value of the work that came before them. They don’t invalidate the careers built within those systems.

Science has always been shaped by cycles of expansion and retrenchment. What endures across those cycles are people—their judgment, integrity, and commitment to evidence—even when institutions falter.

If a rebuilding phase comes, it will depend on that human capital more than on any single appropriation. And if it doesn’t come quickly, the work of making sense of this moment still matters.

Acknowledging the pain, naming the tensions, and refusing to pretend that this is easy aren’t acts of pessimism. They’re acts of stewardship.

I don’t know what comes next for me, or for the many others who have left or will leave. But I know the work we did mattered. The programs we built, the colleagues we mentored, the innovations we championed, the careful governance that made complex research possible—all of that was real, even if it’s being treated as expendable now.

That work still matters. Even in contraction, even in grief, even when the system shrinks around us.

This essay is part of a series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise.

Historical Sources:

1. https://itif.org/publications/2025/06/23/us-science-policy-at-a-crossroads/

2. https://www.ncbi.nlm.nih.gov/books/NBK45556/

3. https://www.historians.org/resource/history-of-federal-science-funding/

4. https://www.nsf.gov/about/history/narrative

5. https://vcresearch.berkeley.edu/news/how-us-became-science-superpower

6. https://ui.adsabs.harvard.edu/abs/2016nsf....1643037T/abstract

7. https://www.journals.uchicago.edu/doi/full/10.1086/588694

8. https://www.congress.gov/crs-product/R43341

9. https://www.nytimes.com/2015/04/22/opinion/double-the-nih-budget.html

10. https://www.clintonfoundation.org/stories/the-golden-years-in-medical-research/

In my first two essays, I focused on the often-invisible roles and structures that shape whether research succeeds or fails. In this essay, I want to turn toward a question that has increasingly defined my later career: why it is so difficult for biomedical research to move beyond silos, even when the need for integration is widely acknowledged.

Few ideas in contemporary health research are as appealing, or as slippery, as “systems thinking” or “whole-person approaches.” We recognize that health is shaped by biology, behavior, environment, social context, and policy. We acknowledge that chronic disease, mental health, and health inequities cannot be addressed one pathway or one intervention at a time. And yet, despite decades of rhetoric about integration, most research remains stubbornly fragmented.

This is not because scientists fail to appreciate complexity. It is because our research systems are still largely designed for a different era.

How Silos Became the Default

Scientific silos did not emerge by accident. They were an efficient response to the challenges of an earlier stage of biomedical science.

As fields matured, specialization allowed for depth, rigor, and cumulative progress. Disciplines developed shared methods, standards of evidence, and training pipelines. Funding structures, peer review panels, journals, and academic departments evolved in parallel. The result was an ecosystem optimized for advancing knowledge within clearly defined boundaries.

That system worked remarkably well for many problems. It enabled breakthroughs in molecular biology, pharmacology, and clinical therapeutics. But it also locked in assumptions about what “good” science looks like: focused questions, tightly controlled designs, and outcomes that can be cleanly attributed to specific mechanisms.

As health challenges became more complex and interconnected, those same structures began to constrain rather than enable progress.

When Integration Becomes Everyone’s Job—and No One’s Responsibility

Integrated research does not fail because it lacks champions. It fails because it lacks infrastructure.

Working across biological systems, disciplines, and sectors requires more than good intentions. It requires shared frameworks, common language, interoperable data, governance models that support collaboration, and analytic methods capable of handling feedback loops and nonlinearity. Most importantly, it requires alignment across funding, review, and training systems.

In practice, however, integration is often layered on top of siloed structures rather than designed into them. Investigators are asked to “add” collaboration without additional time or support. Review panels trained in single disciplines are asked to evaluate work that intentionally resists reduction. Early-career scientists are encouraged to be interdisciplinary while being assessed by metrics that reward disciplinary depth.

The result is predictable. Integrated proposals are seen as too broad, too complex, or insufficiently mechanistic. Promising collaborations struggle to sustain themselves. Systems-oriented work is praised in concept but penalized in execution.

Scale Exposes the Limits of Siloed Thinking

These tensions become even more pronounced at scale.

Large, multi-site programs are often launched to address complex problems, but they are frequently governed as collections of independent projects rather than as integrated systems. Data are collected in parallel but not harmonized. Outcomes are measured inconsistently. Opportunities for synthesis are deferred until the end, when integration becomes both technically and politically difficult.

I saw this occur in national research initiatives. Individual studies were strong, well reviewed, and competently executed. Yet the program as a whole struggled to generate insights across studies that could inform understanding of complex, interconnected systems. The missing element was not effort or expertise, but intentional system-level design.

Without explicit attention to integration, scale amplifies fragmentation rather than overcoming it.

Whole-Person Health as a Systems Problem

The recent resurgence of interest in “whole-person health” reflects a growing recognition that reductionist approaches are insufficient for many contemporary health challenges. But whole-person health is not a single construct or intervention. It is a systems problem.

Studying whole-person health requires methods that can accommodate multiple interacting domains, trajectories over time, and context-dependent effects. It requires data systems that link biological, behavioral, clinical, and social information. It requires governance structures that support coordination across traditionally separate domains.

Most importantly, it requires a shift in mindset. Instead of asking whether a specific intervention works in isolation, systems-oriented research asks under what conditions combinations of factors produce meaningful change. That is a harder question, both scientifically and institutionally.

Why This Is So Hard to Do Well

Moving from silos to systems challenges some of the deepest assumptions in biomedical research.

It complicates causality. It resists clean attribution. It demands tolerance for uncertainty and variation. It often produces insights that are conditional rather than universal.

These features can feel uncomfortable in a culture that prizes control and precision. They also complicate peer review, funding decisions, and publication. Without deliberate effort, systems-oriented work risks being judged by standards it was never designed to meet.

This does not mean abandoning rigor. It means redefining rigor in ways that are appropriate to the questions being asked.

Designing for Integration, Not Just Encouraging It

If integrated approaches are to succeed, they must be designed into research systems from the outset.

That means funding mechanisms that support coordination and synthesis, not just individual excellence. It means review processes that include expertise in systems methods and team science. It means training pathways that equip researchers to work across levels and disciplines without sacrificing depth.

It also means recognizing that integration is itself a form of expertise. Designing governance, data architecture, and analytic strategies for complex systems is not ancillary work. It is central to producing usable knowledge.

I saw firsthand what this kind of intentional design requires when we attempted to build a whole-person research program within NIH.

What Designing for Integration Looked Like in Practice

That gap between aspiration and structure became tangible in my work as a Program Officer at NCCIH, where we were charged with launching a new research program focused on whole person health. The effort required assembling NIH program staff from multiple Institutes, each trained in different disciplines and analytic traditions, and asking them to co-develop a vision for a shared scientific infrastructure and interoperable data ecosystem rather than simply fund individual research projects.

One of the central obstacles was structural. NIH funding mechanisms, peer review processes, and success metrics were still largely optimized for single-discipline hypothesis testing, not for systems modeling or integrative, multi-level research. As a result, multidisciplinary projects often faced a double bind: insufficiently rigorous by disciplinary standards and insufficiently comprehensive by systems ideals.

Addressing this mismatch required selecting an infrastructure-focused funding mechanism, the U24 cooperative agreement, an NIH mechanism designed to support shared research infrastructure, and adopting a phased, milestone-based approach that allowed the science to evolve iteratively over five years in close collaboration with NIH staff. This meant designing governance explicitly, aligning incentives across disciplines, and supporting methodological pluralism while maintaining clear standards for rigor and accountability.

Looking Ahead

The push toward systems and whole-person approaches is not a passing trend. It reflects a necessary evolution in how we understand health and disease. But aspirations alone will not get us there.

As with the failures I described in my previous essay, the barriers to integration are largely structural. They are embedded in how we fund, review, train, and reward science. Addressing them requires intentional redesign, not just enthusiasm.

Moving from silos to systems is hard because it asks us to change not only how we do research, but how we define success. The payoff, however, is the possibility of knowledge that better reflects the complexity of human health and has a greater chance of improving it.

This essay is the third in a series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise.

When people think about how science advances, they usually picture investigators, laboratories, and peer review panels. That picture is not wrong, but it is incomplete. There is another role, largely invisible to the outside world, that quietly shapes what research gets done, how it gets done, and whether it ultimately makes a difference: the program officer.

Program officers do not run labs. We do not publish first-author papers. We do not decide funding outcomes on our own. But over more than two decades inside NIH, I learned that the cumulative impact of program officers’ judgments, often exercised under uncertainty, constraint, and imperfect information, has an outsized influence on the scientific enterprise. Not because of authority, but because of stewardship. Program officers are scientists, advisors, administrators, communicators, and problem-solvers who ensure that research dollars are used wisely, policies are followed, and science advances.

What Program Officers Actually Do

There are a few persistent misconceptions about the role. Program officers do not pick winners behind closed doors, override peer review, or impose pet ideas on the field. Those caricatures miss the point.

What program officers actually do is translate broad public priorities, often set by Congress, the Administration, or emerging public health needs, into fundable and executable science. That translation requires dozens of consequential decisions that rarely appear in public view: which research questions are mature enough to support, which kinds of risk are acceptable, what level of methodological rigor is realistic in real-world settings, and how to structure programs so they can function at scale.

At small scale, science can tolerate a fair amount of inefficiency. At national scale, it cannot. When a program officer is responsible for a $40–50 million portfolio spanning dozens of institutions, design decisions become operational decisions, and operational decisions become scientific ones. Governance structures, data standards, coordination mechanisms, and timelines are not administrative details. They determine whether a program generates knowledge or collapses under its own complexity.

Why This Quiet Role Matters

Because program officers operate upstream, their influence is often indirect but durable. Over time, three levers matter most.

First, methods become norms. When funding mechanisms reward certain designs, analytic approaches, or ways of thinking, those choices ripple outward. They shape what investigators propose, what trainees learn, and what reviewers come to expect as good science.

Second, fields either expand or stagnate. Decisions about who is welcomed into a portfolio, including engineers, data scientists, systems modelers, and implementation researchers, determine whether a field evolves or remains insular.

Third, scale changes everything. Practices that work for a single R01 often fail when multiplied across a national consortium. Without intentional governance, coordination, and shared infrastructure, even well-designed studies can flounder.

I saw these dynamics repeatedly across different Institutes and scientific domains.

Moments Where Quiet Decisions Changed Outcomes

At NIDA, when prevention science was still dominated by relatively static intervention models, we began encouraging adaptive designs and more sophisticated analytic approaches. This meant bringing in statisticians and engineers who did not see themselves solely as “addiction researchers,” and supporting methods that were not yet mainstream. The lag between innovation and acceptance was long, but the payoff was a field better equipped to address heterogeneity, context, and change over time.

At NCI, the challenge was scale. The State and Community Tobacco Control program was not a single study, but a coordinated network of projects, each scientifically strong in isolation. What determined success was not just scientific merit, but whether collaboration was intentionally designed. Governance structures, shared measures, and clear decision rules were essential. Without them, the science would have been fragmented. With them, the program produced integrated insights that no single project could have generated alone.

Later, at NCCIH, I oversaw a large pragmatic trial that was struggling, not because the question lacked importance, but because the original design could not accommodate the challenges of implementation during the COVID pandemic. Adjusting course required methodological realism, careful stewardship of public investment, strategic consideration of grants policy, and respect for participants and investigators alike. The goal was not to rescue a project for its own sake, but to ensure that the knowledge produced would actually be usable.

In each case, the critical decisions were not dramatic or public. They were incremental, technical, and consequential.

The Human Side of Stewardship

Program officers also serve as mentors, translators, and sometimes gatekeepers. We advise early-career investigators when their ideas are promising but premature. We help established scientists rethink approaches that no longer fit the problem at hand. We say “not yet” far more often than we say “yes,” and we do so knowing that real people’s careers are affected.

That responsibility creates an ethical obligation to be honest, to be consistent, and to remember that public research dollars represent trust. That trust comes from taxpayers, from research participants, and from communities who expect that science will ultimately serve them.

A Role Under Strain

Because this work happens behind the scenes, it is easy to underestimate its value. When institutions fail to support or respect program staff, the consequences are subtle at first. Increased risk aversion, loss of institutional memory, and brittle programs that cannot adapt emerge gradually. Over time, those weaknesses accumulate, and the scientific enterprise becomes less resilient precisely when complexity and uncertainty are increasing.

What I Carry Forward

I left NIH with deep respect for what public service makes possible and with a clear understanding of how fragile those possibilities can be. The skills I developed as a program officer, including systems thinking, methodological judgment, governance design, and ethical stewardship, are not unique to government. They are essential wherever complex research is meant to improve real lives.

The quiet power of program officers lies not in control, but in care. Care for the science, for the people who conduct it, and for the public it is meant to serve. That kind of stewardship may be invisible, but its absence is not.

This essay is the first in a series reflecting on what I learned over more than two decades working inside the U.S. biomedical research enterprise.

Day 9. Col de Glandon.

For Stage 19 Jennifer, Joni, and I rode our bikes from our hotel in Les Deux Alpes all the way to Allemont to watch the peloton start ascending the famous Col de Glandon. Along the way we got some terrific photos on the large artwork TDF bicycle displays in Bourg-d'Oisans and near the TDF helicopter site.

Eventually we set up alongside the road near Allemont to wait for the peloton to arrive. The helicopter flew in overhead, the crowd cheered and the racers flew by us.

Then we followed the peloton headed toward the famous Col de Glandon. Col du Glandon (1,924 m (6,312 ft)) is a high mountain pass in the Dauphiné Alps in Savoie, France, linking Le Bourg-d'Oisans to La Chambre. From Le Bourg-d'Oisans the route follows the D1091 through the Romanche valley before joining the D526 after 8 km (5 mi). The climb starts at the Barrage du Verney from where there is a further 24.1 km (15.0 mi) to the summit, which is reached shortly after the junction with the route to Col de la Croix de Fer. Over this distance, the height gained is 1,152 m (3,780 ft); the average gradient is 4.8%, although there are some downhill sections en route and a maximum uphill gradient of 11.1%.  Our plan was to ascend the Col de Glandon and meet up at the lone restaurant for lunch and then continue on to ride the Col de la Croix de Fer as well. Jennifer gave me a few Euros since I was likely to get to the restaurant Chalet Hôtel Col du Glandon first and we all set off.

As I started up, there was a slight drizzle that came and went, but it did not get me wet or pose a problem for the climb. I relaxed into a very fine tempo which I was able to maintain all the way up into the small town of Le Rivier d'Allemont. There was a man riding directly behind me drafting my wheel for the last few kilometers up to the town. When we crested the top he pulled ahead of me and thanked me for the pacing. He said it was perfect! I laughed and asked him if I was his domestique. He smiled and laughed back. The town was small and very scenic so I pulled over to snap a few photos before heading on. At this point I think I might have been at about the halfway point to reach the Chalet Hôtel Col du Glandon.

There is a slight downhill at this section here with several switchbacks that take you all the way down into a valley and from there you begin the second part of the long ascent. I settled back into my rhythm and tried to keep a descent pace. There were lots of other cyclists making the trek up, but the road was not crowded and everyone just went at their own respective pace. The views are stunning as you ascend through the mountain pass with streams rolling by and you end up at a glacial lake way up high. This goes on for what feels like forever...

I arrived at the restaurant at 1:30, parked my bike, and sat down and ordered lunch seated outside on the patio. It was chilly but clear and pleasant. Lunch was terrific: homemade vegetable soup with croutons and cheese and a lovely green salad.

By 2:15 I began to wonder if Jennifer and Joni were coming. The temperature was dropping and the skies started to darken. By about 2:30 it started to rain, and then it started to pour. The temperature started to drop even more dramatically. All of the cyclists seated on the patio including me piled into the small restaurant building to stay warm and dry. I thought it would be good to wait out the storm rather than try to descend in it, but the rain just did not let up. Three o'clock came and went. It was getting late. The only thing that time in the restaurant bought me was more drops in the temperature. Now you could see your breath in the air when you exhaled. It was feeling like winter--very cold, probably close to 40 degrees--and I was dressed for summer. Not good. I didn't know this at the time but this same storm had rolled in over Jennifer and Joni earlier on the climb and it was bad enough to cause them to abandon the climb just after Le Rivier d'Allemont. They turned around and had returned to Allemont to wait for me there. Unfortunately my cell phone was useless up on the top of the Col du Glandon so I had no clue where they were or what happened.

There were several bike tour groups at the cafe taking refuge with me. There must have been at least 40 cyclists in the cafe. Almost all of them were with big tour companies and they had sag vans waiting to take them back to their hotels in the storm. Most were planning to eat lunch and then just hop in their vans. However, none of these vans were going in my direction, and besides, they were all full. I spoke to one of the guides to get some advice. It was still pouring, the temperatures were dropping, and the descent down the switchbacks of the Col back to Allemont would be quite treacherous. This tour guide was advising his own folks to get into the van for safety. I explained to him that there was no van coming for me and that I really had no options to get home except on my bike. Fortunately I had my backpack with me and it still contained all of the extra clothes and gear that I had intended for emergency weather on Alpe d'Huez. I remembered that I had extra socks, long fingered gloves, toe warmers and a rain jacket! I put on every piece of clothing that I had with me: toe covers, arm warmers, vest for core warmth, and a light jacket on top of that, long gloves, and the clincher, my Eddie Bauer red rain jacket with the big hood which I pulled up and over my bicycle helmet. With all of this gear on the tour guide looked at me from head to toe, shrugged, and told me that I was as ready as I'd ever be. He advised me to leave immediately or else I would risk even more temperature drops. I was going to have to descend the mountain pass in this storm on my own. To tell you the truth I was terrified to descend in the pouring rain all alone. What if I crash and lie injured on the road? Would anyone find me?

I noticed that there were two Belgian cyclists in the cafe that were not with any of the tour groups. They were on their own and in the same boat as me and were also leaving. I quickly jumped on my bike and tried to follow them. One of the Belgians was more confident on the slick steep roads than the other one. He led the way and at one point he got too far ahead of his buddy so he stopped to wait. I caught up with him and asked him if they would mind if I descended with them. I explained that I was all alone and that my cell phone didn't work and that my plan was to just get farther down the road back to Le Rivier d'Allemont to send a text message to my tour group. He was very kind and agreeable, so we set off together. The confident one was in front; I took the middle; and the other (more tentative) fellow took up the rear. It was pouring rain. Visibility was low and the puddles were large. When cars passed us we were forced into the puddles and were sprayed. I was terrified of my brakes failing on the steep grades or of fish-tailing my wheels on the slick pavement. We went real slowly down the mountain--about 9-10 mph. It took a long time and a lot of concentration. Once we passed the steep switchback part down into the valley we had to traverse two flooded sections of the road. We tried to pick the best part with the lowest water line. After we made it through the submersed sections of road we were finally at the slight ascent up to the village town Le Rivier d'Allemont.

I started up the switchback ascent and I slowly pulled ahead of the Belgians on the climb. Soon I could not see them behind me anymore. They were much slower than me on the climb. At the top I pulled up next to a cafe in Le Rivier d'Allemont and checked my cell phone. Still no service. But all of a sudden the rain stopped and the sun came out! Just like that. The roads were still wet but the storm had finished here. I pulled off my drenched long gloves and pulled my short dry ones out of my pack. I sent Jennifer a text message to let her know where I was and I got ready for the final descent into Allemont. Just as I was about to pull out again the two Belgians rode by and waved. They had finally made it up the switchback climb section and would you believe it, they pulled over for a smoke! Seriously? Here? We’re riding one of the most challenging cycling routes in the world and they are smoking? Clearly this was not helping their climbing. No wonder I beat them handily on the climb. I waved back, wished them well, and I made my way down alone the rest of the way to town. The road was damp, but no more rain. When I finally arrived in the valley I removed my rain coat and put it back in my pack. The wind on the way down had completely dried my rain coat, but my feet and shoes were soaked. It was nice and warm down in Allemont and I finally had cell phone service. Several backlogged texts from Jennifer started coming in as my phone dinged furiously at me.

In a series of frantic text messages from Jennifer I learned that they had turned back when the rain hit hard and they waited in Allemont for me the whole time, but they finally gave up. At this point they were well past Bourg d'Oison near the turn off choice point for Les Deux Alpes and Venosc and headed home. I quickly texted them and they said they would wait for me near the turn off to Venosc. After a quick bathroom break I rode as fast as I possibly could now as it was mostly flat terrain. The aluminum rental bike with a 50t big ring just did not have the get up and go power of my carbon Trek madone with a 53t big ring, so my little legs couldn't get the bike to go much faster than about 21 mph. Nevertheless it felt like a time trial. By the time I arrived at the meeting place and found Jennifer and Joni the skies were growing dark again. The rain was starting. We all put our rain jackets back on and then it started to pour. We had a 7 km ride up to Venosc to catch the gondola back up to Les Deux Alpes and it was pouring for the entire climb. Eventually we arrived at the top of the mountain via the gondola and we rode to our hotel. I was shivering and my teeth were chattering. I couldn't wait to get there. I was completely soaked. 67.5 miles on my bike with over 8400 feet of climbing. The hot tub that night was awesome.

After we warmed up and took our showers we headed out for a celebration dinner at Casa Nostra in Les Deux Alpes. This is an excellent restaurant! I highly recommend it if you ever travel there.

Daily Ride data:

On Strava http://app.strava.com/activities/69110151

Day 8. Alpe d'Huez.

In the morning I packed my backpack full of extra food and extra clothes just in case it might be cold up on the mountain. Rain was also a possibility, but we were not going to miss Alpe d'Huez! I left the water pack portion of my backpack at the hotel as it was not going to be hot like Ventoux and this allowed me to fit more clothes, supplies, and a rain jacket into my pack.

We rode our bikes down from Les Deux Alpes in the morning so that we could ascend Alpe d'Huez with the massive crowds. Alpe d'Huez was a huge party on wheels (and beer)!

There were big bottlenecks at times with hundreds of riders ascending at once so we had to walk our bikes a few times. The climb goes via the D211 from where the distance to the summit (at 1,860 m (6,102 ft)) is 13.8 km (8.6 mi), with an average gradient of 8.1%, with the 21 famous switchbacks all numbered as you ascend and a maximum gradient of 13%. L'Alpe d'Huez is climbed regularly in the Tour de France. It was first included in the race in 1952 and has been a stage finish regularly since 1976. The fastest recorded professional times up were just over 37 minutes, but these cyclists were later found out to have been doping! Greg LaMond the famous American cyclist who did not dope made it up in about 48 minutes. Depending on what websites you believe the American musical icon Sheryl Crow reportedly climbed ADH in about 1 hour 29 minutes to 1 hour 40 minutes when she was dating Lance Armstrong. I don't think he shared his EPO with her; she did it on her own natural ability. However, I'm quite certain he gave her a $12,000 carbon Trek bike to use for the ascent.

As I was riding up Alpe d'Huez there were professional photographers snapping riders' photos. One of them handed me a card as I ascended so I was able to order my shots from their website. A fun bit of memorabilia! Just after I reached the 4km to go banner the road split and the gendarmerie (police) forced me to ride up the remainder of the way on the road to the right. This is not the official way; it's called the "alternate route,” but I can assure you that it was no less steep. I got to top on this side in about 1 hour 15 minutes. Not bad! At least I beat Sheryl Crow. After looking around a bit and snapping some photos I descended back down to the 4km to go sign so that I could attempt to go up the "correct" side of the mountain. So basically I did the top of ADH twice! What's better than one sufferfest? Two!

Once I was done with my double ascent I descended on my bike to switchback #5, our designated meeting spot, to try to meet up with my riding companions. Only Chuck made it. Jennifer and Joni were stopped at the summit and had to view the race from up there, but we would not know this until later that evening. Chuck and I stayed at switchback #5 and partied with a cool bunch of Brits in blue Union Jack body suits. Allez!  It was fantastic. We did the wave, cheered for the amateur riders going up, sang to the music and thoroughly enjoyed the party atmosphere. We even ran into Lesli, Dave, and Justin from Cyclismas who were also partying Alpe d'Huez style! Lesli told us that Dave and Justin were trying to figure out how to get a keg up to the summit on a bike. Basically it seemed like everybody except the children were drunk on Alpe d'Huez.

Once the race was on we had twice the fun because the riders had to ascend Alpe d'Huez twice for the stage. We saw Teejay Van Garderen lead the pack up the mountain two times! So exciting. However, Van Garderen's efforts were unfortunately in vain as the Frenchman Christophe Riblon overtook him in the final 2km for the win.

One interesting calamity occurred for me and my riding companion, Chuck. We had fully expected to meet Jennifer and Joni at switchback #5, but they were not allowed to descend from the summit. Before the race Chuck and I had found a place near the woods to stash our bikes while we watched the stage. We had a bike lock with us, but Jennifer had the key. So I advised Chuck to lay simply the lock over the bikes and make it appear that they were locked, but instead he actually locked it! Our bikes were secured together with no key. With nothing to do about it we watched the race and hoped that the key would come. However, once the race was over it became clear to us that no key would be arriving. How on earth would we descend with our bikes locked together? No problem. It turns out that there a lot of great would-be bike thieves up on Alpe d'Huez. In no time at all Chuck was able to find a Belgian with a small hand saw and a Dutch guy with drill. They had a drunken contest to see which one of them could break through our bike lock first. The Belgian won hands down and we were on our way.

The descent down Alpe d'Huez post-race was challenging to say the least. Imagine thousands of spectators on foot or on bikes along with hundreds of cars and vans. Add to that the entire procession of the TDF caravan which had been parked up on top after the first ascent and needed to come down. We were all packed onto one narrow, steep road and made our way down the 21 switchbacks together.  What a nightmare. It felt like two days of descending and my hands were on fire from clenching my breaks, but we finally made it down. Next up: how to find Jennifer and Joni. They took a different road for the descent in the hopes that it would be less crowded. No such luck. Their route was just as gridlocked as ours and they faced a very local flash thunderstorm that only hit that side of the mountain. They arrived completely soaked. Once together we all rode back to the van in time to drive home for a late dinner.

Daily Ride data:

On Strava http://app.strava.com/activities/69110150

Day 7. Travel Day.

We packed up in the morning and after breakfast loaded up the van and headed toward Les Deux Alpes. We originally planned to do a ride near the hotel near Violès first, but it was hot with a threat of thunderstorms, and frankly, we were all very tired from the previous three days of riding with not enough sleep. Collectively we decided to skip riding and take some recovery time so that we would be ready to ride the big mountain the next day.

Our drive to Les Deux Alpes from Violès was expected to take about three hours with some stops for shopping, pictures and food along the way. Jennifer suggested a detour that would extend the travel day but would allow us to drive through the Gorges in the Vercors Regional Natural Park on our way toward Grenoble. This is a protected area of forested mountains in the Rhône-Alpes region of southeastern France. The views were absolutely breathtaking as we rode in the van through the Gorges de la Bourne and on through Grenoble. Did I mention that the food in France is incredible? Even at the highway rest stops you can purchase fresh gourmet food such as homemade tortellini with fresh basil pesto!

We arrived at our hotel in Les Deux Alpes, the Hotel Le Souleil’Or, in the early evening hours.  www.le-souleil-or.fr  It was located in a very beautiful setting high up in the mountains with Alpes style architecture which reminded me of Austria. I had a slope side view from the hotel room. There was no snow on the low slopes but folks were are still skiing up top on the glacier this time of year. Down below near our hotel the ski trails were filled with cows grazing on the grass by the chairlifts. You could hear their cow bells gently ringing in the wind. Naturally I ordered raclette for dinner at Le Charbon de Bois Restaurant.

We would do Alpe d'Huez tomorrow with a forecast for chilly temps and even rain possible.

Day 6. Stage 16 Vaison la Romaine. Fun to watch the boys roll out!

On Tuesday we rode our bikes from our hotel in Violès to Vaison-la-Romaine to view the start of Stage 16. We saw the racers at the ceremonial podium sign in and we saw them depart for the stage. It was a very festive scene!

Afterward we did a really nice rolling ride around the vineyards. We had lunch at the street market in Vaison-la-Romaine overlooking the sites of Roman ruins. The vendor food at the market was absolutely incredible. We purchased fresh roasted chicken, several different kinds of olives, roasted red pepper, and a fresh pineapple for dessert. We ate sitting on the sidewalk in our cycle clothes/shoes with the passerby greeting us, "Bon appetit! Bon appetit!" Unbelievable. Sure beats my usual ride stops in the states at the Sheetz and the High's convenience stores.

After lunch we climbed to Suzette where we had a cool beverage at a cafe and then we rode over to Gigondas before heading home. In the evening we attended a really nice free wine tasting in Gigondas and then we had dinner there (which in France always takes 3-4 hours). There is wine and great food everywhere. All amazing.

Daily Ride data:

On Strava  http://app.strava.com/activities/69110115

Day 5. Gorge de la Nesque and Mont Ventoux again.

It was a rest day for the TDF riders, but not for our travel group! We did an epic 47 mile ride starting in Villes-sur-Auzon and winding our way through the Gorges de la Nesque and through miles of very beautiful and fragrant fields of lavender on our way to Sault for a lunch rest stop. Riding along the gorge was a real treat as this was one of the most scenic routes I have ever had the pleasure to ride. It was dreamlike.

Then riding through the lavender fields on the way to Sault was absolutely beautiful. The aroma was truly AMAZING.

In Sault we had a lovely lunch at a French cafe with a panoramic view of Mont Ventoux: the restaurant O'PICHOUN à Sault. Nearly all the patrons in the restaurant were cyclists taking a break so we felt right at home. Many of them were carbo loading plates of pasta in preparation for the ascent of Ventoux.

Après lunch it was back on the bikes in an attempt to redo Mont Ventoux which we were unable to complete the day before. This time we went up Ventoux from the other side starting in Sault. From this direction the route is supposedly a little easier (but not too much easier, mind you!). It was a similar distance -- 21 km.

I made it to the top of all 21 km in approximately 1 hour 45 minutes. I rode a much stronger, faster pace than the day before without the crowds. This time I had only had one compulsory stop on the way up as I encountered a very large herd of sheep crossing the road near the summit.

The summit of Mont Ventoux is legendary. As you arrive it feels sort of magical like you have just won an imaginary Olympic medal or something. People clap and congratulate all of the riders as they arrive and they take turns taking their photo underneath the famous Ventoux summit sign. The views are stunning but also bizarre and somewhat surreal. Originally forested, Mont Ventoux was systematically stripped of trees from the 12th century onwards for shipbuilding. The top is essentially barren with windblown, sandy rock. It looks like you are on the moon or perhaps on a Star Wars movie set. I've seen it on TV in previous years of the TDF, but seeing it in person was definitely worth the climb.

The descent from the summit of Ventoux down to Malaucène took about 40 minutes and it was very cold (Brrrrrr!). This was the longest, most epic descent of my life! No stops this time, no crowds, just open road descending with a few cars here and there. That night we had a wonderful dinner in Malaucène. However, unfortunately during dinner Jennifer received a phone call and we learned that our new friend Allistair from Britain had come down the same descent from Ventoux alone and he blew a front tire at high speed and crashed out. He fractured his collarbone in three places and spent six hours in the hospital trying to be treated. Late that night Jennifer retrieved Allistair from the hospital and brought him back to the hotel. He left the hospital with nothing but an x-ray picture and a shoulder brace. He would need to seek surgery back home in Great Britain.

Daily Ride data:

On Strava http://app.strava.com/activities/69110147