Send my tax dollars
to Mississippi

Published October 03 2016

Taxpayers fund the National Institutes of Health, the NIH funds biomedical research, and benefits of that research are returned to taxpayers. Understandably, NIH officials, such as Deputy Director for Extramural Research Michael Lauer and Director of the National Institute of General Medical Sciences Jon Lorsch, are interested in maximizing the return on taxpayers’ investments (1, 2).

One key challenge lies in the fact that NIH funding is allocated disproportionately to a minority of investigators, institutions and states. Lorsch, Lauer and others point out that these skewed distributions of funding lead to diminishing marginal returns on taxpayers’ investments (2–6).

In Lorsch’s example, using NIGMS funding data, $200,000 annual direct costs for a first R01 grant, such as one to a new principal investigator, would, on average, buy the taxpayers approximately five scientific publications during the funding period. Remarkably, the same amount of funding for a third R01 grant to an established investigator would buy the taxpayers, on average, only one additional publication.

“The choice seems obvious,” said Lorsch in a 2015 piece he wrote for the journal Molecular Biology of the Cell (2). “Taxpayers net four more papers by funding the new PI than by giving the established PI a third grant.” Disparities (or biases) in the allocation of funds to individual investigators can undermine the productivity of the nation’s biomedical research enterprise.

The same principles apply at the level of institutions. Let me illustrate this by comparing the five top-ranked institutions to 10 arbitrarily chosen, lower-ranked institutions listed in the 2016 U.S. News & World Report rankings of “Best Medical Schools: Research” (7). The data, which are available to the public, are total research project grant, or RPG, funding (8) and RPG-supported scientific publications from 2006 to 2015 (from PubMed).

Figure 1. Grant funding and productivity by institution (2006-2015). Institutions are listed by rank according to “Best Medical Schools: Research” (7) (RNP, rank not posted; UR, unranked). Amounts of NIH research funding and numbers of grant-supported publications are from NIH RePORTER (8) and PubMed, respectively. FIGURES BY WAYNE WAHLS

Each of the five top-ranked institutions received more RPG dollars than each of the lower-ranked institutions (Figure 1). This makes sense, given that amounts of NIH funding (total and per faculty member) were criteria used by U.S. News & World Report for rank ordering. Notably, each of the lower-ranked, less-funded institutions produced more publications per dollar of RPG funding than each of the top-ranked, highly funded institutions. Plotting the data a different way reveals diminishing marginal returns on investments relative to total funding, to mean funding per project and to mean funding per principal investigator (Figure 2).

The choice seems obvious: Taxpayers net more scientific publications by funding investigators at the University of Mississippi Medical Center (and other low-ranked institutions) than by giving the funds to prestigious and top-ranked institutions.

Figure 2. Relative returns on taxpayers’ investments. Plots show publications per dollar of NIH research-grant funding as a function of (A) total funding, (B) mean annual funding per project, and (C) mean funding per principal investigator at each institution. Lines and statistical values (inset) are from linear regression; the curvature in panel A is due to plotting total funding on a log scale.

Why are quality rankings and funding allocations (to individuals, institutions and states) discordant with productivity metrics? I suspect the answer has to do with implicit bias during the allocation of funds (9). Scientists and NIH officials who review grant applications are influenced by pervasive subconscious attitudes or stereotypes that can differ substantially from quantitative realities. Even quantitative realities can be misleading. For example, perceptions about the quality of institutions based on total research funding or funding per faculty member (bigger must be better!) are flawed because they fail to normalize for the number of faculty members actually doing research. More fundamentally, such metrics provide no insight into return on investment.

We cannot eliminate subjective assessments central to grant review, journals’ decisions on which manuscripts to publish, and authors’ decisions on which papers to cite. We cannot avoid the implicit biases and overt perceptions that shape our subjective assessments. But we can measure and adjust for disparities and biases in allocation and outcome that stem from our subjective assessments. Cogent arguments for optimizing the allocation of funding at the level of investigators (1, 2, 5, 6) apply equally well at the level of institutions (10) and states (11). Such adjustments would mesh nicely with, and should be a key component of, NIH initiatives to address institutional and geographical funding bias and to promote the diversity, productivity and sustainability of the nation’s biomedical research enterprise.

We need systematic analyses of funding versus productivity differentials by institution and of how those values compare to grant application success rates. Well-funded institutions, like well-funded investigators (12), should receive extra scrutiny. Meanwhile, I encourage the NIH to invest a greater fraction of my tax dollars in places like the University of Mississippi Medical Center, because these low-ranked institutions can provide greater returns on taxpayers’ investments than prestigious institutions that currently receive a disproportionate share of NIH research funding.


REFERENCES

  1. Lorsch, J.R., iBiomagazine (2015).
  2. Lorsch, J.R., Mol. Biol. Cell 26, 1578 – 1582 (2015).
  3. Doyle, J.M., et al., Mol. Phychiatry 20, 1030 – 1036 (2015).
  4. Berg, J.M., NIGMS Feedback Loop Blog. (2010).
  5. Fortin, J.M. & Currie, D.J., PLoS ONE 8, e65263 (2013).
  6. Gallo, S.A. et al., PLoS ONE 9, e106474 (2014).
  7. US News & World Rep. (2016).
  8. National Institutes of Health Research Portfolio Online Reporting Tool (RePORT). (2016).
  9. Kirwan Institute for the Study of Race and Ethnicity (2016).
  10. Murray, D.L., et al., PLoS ONE, 11, e0155876 (2016).
  11. Wahls, W.P., Peer J. 4, e1917 (2016).
  12. Berg, J.M., Nature 489, 203 (2012).
Wayne Wahls Wayne Wahls is a professor of biochemistry and molecular biology at the University of Arkansas for Medical Sciences. His research on meiotic chromosome dynamics is supported by grant GM081766 from the National Institutes of Health.