Reforming the NIH’s Intramural Program

A guest post from Christopher Marcum

Christopher Steven Marcum, PhD, FGSA¹
ORCID: https://orcid.org/0000-0002-0899-6143

In this guest post, Chris Marcum (a friend of the Good Science Project) reflects on his experience with the Intramural Research Program (IRP) at the National Institutes of Health (NIH), and makes some needed suggestions for reform.

Summary

The Intramural Research Program (IRP) at the National Institutes of Health (NIH) needs substantial overhaul and reform. Its original mission was to support “high-risk, high-reward” research without the need to compete for university positions and external federal grants. Yet, with a few notable exceptions, the IRP has increasingly become more akin to the competitive grant-funded research conducted at a university. Suggested reforms include better data collection and transparency, and refocusing the entire program on truly outside-the-box research.

The Challenge

The National Institutes of Health (NIH) is a behemoth of a federal agency. With recent budgets in excess of $47B, NIH annual expenditures often exceed the total value of the endowments held by private biomedical research institutions like Harvard University and Howard Hughes Medical Institute. Most of that budget is spent on funding the extramural research program. That’s the program that provides grants to fund proposals that researchers at universities and other institutions compete for and has been the lifeblood of biomedical innovation for the last 78 years in the United States.

Perhaps less well-known, however, is that NIH also conducts its own research through the Intramural Research Program (IRP). Around 10% of the total NIH budget is spent on intramural research. In FY 2023, that amounted to more than $4B in support for research conducted at the NIH itself. The IRP is a natural outgrowth of the NIH’s historic legacy of starting out as primarily an intramural research institution–that is, federally-funded biomedical science was conducted solely within government labs between 1887 (e.g. at the Marine Hospital Laboratory of Hygiene) and 1946 when the external grants and fellowship program was established.

This seemingly meager investment of just around the historical average of 13% of all research spending has had an outsized impact. Many groundbreaking discoveries have happened in the NIH Intramural Research Program. In the heyday of the 1960-70s, NIH employees (such as Marshall Nirenberg and Julius Axelrod) won four Nobel Prizes within an eight-year period! The Division of Intramural Research at the NIH boasts that its “unique funding environment means the IRP can facilitate opportunities to conduct both long-term and high-impact science that would otherwise be difficult to undertake.” This is an environment that, at least in theory, seeks to conduct “high-risk, high-reward” research over a longer period of time than would be feasible under the competitive extramural grant program. There are only a handful of similar schemes available to extramural researchers, mostly funded through special competitions of the Office of the Director’s “Common Fund.” Recent

Competition in science is not necessarily productive, however. A recent paper by Schweiger et al. examined the costs of competition for research support from funders around the world like the NIH. They found that competitive systems are strongly biased against high-risk, high-reward research largely due to the myopic criteria and inherent biases that peer-reviewers use when evaluating proposals. One would presume that this is precisely why the IRP exists: to fill the gap created by the extramural program that values sure success over innovation.

Despite the outsized potential of the NIH’s intramural program, its recent history has been less impressive. For instance, compared to the four Nobel Prizes won between 1968 and 1976, only two NIH employees have won Nobel Prizes in the nearly 50 years since that time (more on NIH Nobels here). Time will tell whether the relatively few big wins for the NIH IRP in more recent years - such as the break-neck speed at which Kizzmekia Corbett and Barney Graham were able to develop proteins for use in COVID-19 vaccines in 2020, or how John Kash et al. were able to track daily changes in Ebola gene expression in a patient as they recovered at the NIH Clinical Center in 2015 or how - will eventually lead to more Nobels for the IRP. I suspect, however, that the recent scaling back of truly high-risk, high-reward research in lieu of a focus on more traditional research means that the years of back-to-back Nobels are behind the IRP.

My own experience is illustrative.

I spent nearly a decade working for the intramural research program as a staff scientist.² As I observed, the “competition” to pursue various research ideas effectively happened at the hiring level, because once you were hired into the tenure track you were given a large budget to do effectively whatever you wanted to do, however you wanted to do it (within the limits of integrity, of course). The hiring process is rigorous–and while it brings in the very best people–it’s not the most transparent process for ensuring that the actual science that those people would conduct could not otherwise be done outside of government. Each investigator, and each institute's IRP portfolio overall, is reviewed every four years by the external constituents of the Board of Scientific Counselors.

The NIH states that: “Rigorous reviews are critical for planning and to ensure that only the most outstanding research is funded. If a program fails to meet expectations, funds are redeployed.” I went through two of these reviews when I was at NIH - they are tough - but tough in a bureaucratic sense not in a competitive sense. A full year is spent preparing for the reviews, mostly by assembling research portfolio summaries, updating biosketches, writing impact statements, practicing talks and making research posters, etc. The reviews are entirely retrospective over the previous three years of work - they do not, for the most part, evaluate prospective work (though the reviewers can recommend it). As such, investigators are effectively competing against themselves to demonstrate progress (even if their research requires more than four years to generate data).

For review of specific research proposals, each institute has a Scientific Review Committee (SRC), which ostensibly does the functions of the extramural equivalent competitive proposal review study panels through the Center for Scientific Review. The committees are composed of NIH career staff. I was a member of the NHGRI SRC for a few years. During that time, I cannot recall a single proposal I ever saw as a member of that committee that I thought couldn't be done at least as well - and perhaps much better - at a university. Investigators often dismissed reviewer comments that could have improved the study without significant consequences (especially my comments, which came from a statistical and social science perspective given to mostly physician-researchers who devalued such). I cannot recall ever seeing a research proposal outright rejected by the SRC.

Minimally, in my opinion, proposals that were not “high-risk, high-reward” research should have been tossed out by the SRC to meet the mission of the IRP. I'm sure many proposals rise to that level somewhere in the IRP but I did not see them on my desk. Because the SRCs are staffed by peer colleagues of the investigators from within the agency, there is a perverse incentive to keep on keeping on without delving into mission-specific politics at that level.

Government in - Government Out

With a budget of just about $800K a year, the IRP was a windfall for the lab I managed. We were a juggernaut at producing outputs in the bio-behavioral research space - my publication record alone was more than double what I could have accomplished as a tenure-track professor at a university with an R01 grant in the same amount of time. And I’m incredibly proud of that work and the work of my colleagues. Still, even though I was able to produce more scholarly outputs than I would have under the support of the extramural program, I now believe the substance of the work would have been achievable at a University without the IRP banner.

Because the IRP is within the Federal government, it has its fair share of waste. Every fiscal year, labs rush to spend down the budget they did not use on research – buying unnecessary equipment, hiring contractors for imagined work, etc – just so that the obligated funds do not disappear in the next year. I once heard about a lab that purchased a $250K microscope right before the end of the fiscal year in 2013 when funds would have disappeared, only to have that microscope become lost upon delivery because of the subsequent government shutdown at the start of fiscal year 2014.

And there is a mountain of bureaucracy that researchers have to climb to manage their projects - government procurement, hiring, and paperwork rules all apply here. Schweiger et al. (ibid) found that competitive grant programs can spend over 50% of their budget reapplying for additional grants and complying with administrative burden. In a non-competitive space like the IRP, one might assume that the paperwork burden would be low. Alas, the last place anyone should look for efficiencies in paperwork is inside the Federal government.

I suppose one saving grace for the IRP is that there was an exemption to the Paperwork Reduction Act passed under the 21st Century Cures Act that meant our research surveys did not have to go through external review by OIRA. Survey research is critical for much social and behavioral work done in the IRP. The downside of that exemption, however, is that there is no public oversight, no coordination with existing survey collections, and no assurances that research surveys rise to the OMB collection standards. I’m confident that additional oversight would improve the quality of these collections, ultimately improving the integrity of research and reducing burden on respondents.

After reflecting on my experience in the IRP, I walked away not fully understanding why federal employees in the NIH intramural program should effectively have a near-monopoly on (the potential to do) high-risk high-reward biomedical research, especially when their advantage is accompanied by less public oversight and transparency. Even worse, I am not fully convinced that the IRP is actually doing high-risk research across its full portfolio. Perhaps, over time, research programs within the IRP have grown more similar to research programs in the extramural community.

The NIH itself appears to acknowledge that the extramural research community is increasingly capable here as its funding support for the IRP has waned over time. While historical data on expenditures in specific extramural versus intramural programs going back to 1946 is sparse, the NIH has kept and posted excellent records since 1983. Since then, in constant 2023 dollars, the median annual percent of NIH intramural research spending was about 13% of the total research expenditure.

As NIH spending on the extramural research ballooned over that 40-year-period, the share of research expenditures on the IRP declined over time to about 12% from its high of nearly 17% in 1983. Does this shift in allocation mean that the relative value of the IRP is declining over time? Are extramural researchers, perhaps those supported by other funders like CZI, HHMI, and Gates, doing more of the high-risk, high-reward science that the IRP champions as its domain?

The Solution

Calls for reform of the IRP have been happening for at least 15 years, such as the 2009 FasterCures Task Force report on the IRP.³ As policymakers consider reform of the NIH in the coming months and years, they should take a hard look at the NIH IRP. Several reforms would help:

• Transparency and Data. Currently, there is very little public transparency or external oversight/evaluation with regard to IRP research programs and expenditures. NIH should allow more external scholars to have access to the internal data that would be necessary to compare the productivity and characteristics of IRP research to externally-funded research.

• Decreasing Bureaucracy. We know that university scientists report spending inordinate amounts of their research time on bureaucracy and administration. More transparency about IRP researchers and how they spend time would likely reveal similar burdens. A focused effort to lower the administrative and bureaucratic burden of NIH IRP investigators would free up time and resources that could be spent on doing actual research.

• Flexibility. One simple policy recommendation would be to allow for NIH IRP to have funding carry-over from one year to another, removing the perverse incentive for internal labs to spend-down on non-necessities every fiscal year.

• High-risk, high-reward research. Due to lack of data, no one really knows the landscape of truly high-risk, high reward research carried out within IRP compared to the broader research community. Understanding whether the IRP retains its competitive advantage requires an evaluation of every single ongoing and proposed project and comparing each against the body of work happening elsewhere. If the goal of the IRP is to conduct work that cannot be done outside of the NIH, then IRP projects that are demonstrably not unique to the program should be excised.

  • For process reform, NIH could experiment with mixing IRP research proposals into the extramural review process with additional criteria about risk and reward: Is anyone already doing the work or could anyone outside of NIH do the work? Are the chances of success sufficiently low to be considered high-risk, and would the success of the research result in substantial benefits to society?

• Replication. The incoming presumptive NIH Director Jay Bhattacharya is already thinking about this with proposals to focus on research replication. Turning part of the IRP into a replication incubator could involve taking a random sample of published results from extramural and intramural research projects and using the IRP resources to evaluate the extent to which NIH funded research is, in fact, genuinely replicable. Such a service would benefit taxpayers by providing insight into the reach of their investment, reveal where greater transparency is needed in data and methods of published research, and assist researchers in identifying potential holes in their workflows.

• Translational and Clinical Work. The FasterCures report proposed that the NIH IRP should make a more consolidated effort to take advantage of the NIH Clinical Center to attack high-risk translational areas (where there is often a “valley of death”). Moreover, the IRP and the Clinical Center should work more closely with FDA to “address urgent questions pertinent to emerging diagnostics, devices, therapies, and vaccines.” Tellingly, these reforms look a lot like steps taken that led to the successes we saw come of the IRP during Ebola and COVID. The full IRP program should do so all the time, not just in temporary emergencies.

1. Dr. Marcum is writing in his personal capacity and the opinions and ideas expressed in this work are solely his and may not necessarily reflect those of his employer or associates.

2. Initially hired as a non-tenure track scientist at NHGRI in the Social Network Methods Section of the Social and Behavioral Research Branch, I made several attempts to switch tracks during that time before I effectively gave up and took a policy position at NIAID helping to lead the Office of Data Science and Emerging Technologies in the Office of the Director during the pandemic.

3. This was brought to my attention by Michael Stebbins, whom I thank.