The world’s largest funder of biomedical research, the US National Institutes of Health (NIH), has decades’ worth of data on grant applications, peer-review results, funding outcomes and publications. Using such data to guide NIH policy could offer insights and help to address inequities. It would align with the “radical transparency” promised by the US health secretary, Robert F. Kennedy Jr, under whose purview the NIH falls.
Grant proposals drafted with AI help more likely to win NIH funding
During a visit to the Broad Institute in Cambridge, Massachusetts, in December last year, NIH director Jay Bhattacharya said, “I want Iowa, Nebraska scientists, scientists at every institution, to be able to compete on the same level playing fields with the brilliant scientists here in Massachusetts.” He said something similar at a hearing of the Senate Committee on Health, Education, Labor and Pensions on 3 February.
But is the playing field really not level? What drives the distribution of NIH funding across the United States?
Information about the number of grant applications and the likelihood of them being funded as a function of applicants’ geography has not generally been made available — even though the distribution of NIH funding across states and organizations is accessible (see go.nature.com/4adpitd). An exception is information on grants that offer small (mainly biotechnology) companies a way to obtain funding for early-stage research and development. Data for these grants — called Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) — are publicly available for the fiscal years 2015 to 2024 (see go.nature.com/3zz8a6z).
An analysis of these data reveals that Massachusetts did do better than Iowa and Nebraska when it came to grant success rate. But the differences were not statistically significant — at least in this data set, although they might be in a larger one. Almost 20% of grant applications were successful in Massachusetts, compared with around 18% in Iowa and around 13% in Nebraska. (Statistically, P = 0.75 for the difference between Massachusetts and Iowa; P = 0.44 for the difference between Massachusetts and Nebraska.)
More importantly, further analysis shows that it was mostly the number of applications submitted by each state that drove the amount of funding each state received (see ‘Applications drive awards’).
Source: NIH
It might seem shocking that California received a whopping 21.7% of the nation’s SBIR and STTR grant funding and Iowa received only 0.56%. That is, until one realizes that investigators in California submitted 21.4% of the applications, whereas those in Iowa submitted only 0.55%. California has built its tremendous research capacity through state investments in higher education and through establishing connections between universities such as the University of California, San Francisco, academic institutions in the San Diego area and private-sector entities connected with research, including many biotech firms.
How Facebook, Twitter and other data troves are revolutionizing social science
This basic pattern is likely to apply to other grant types, too, including the NIH’s bread-and-butter Research Project (R01) grants. The numbers of R01 applications have not been made publicly available. But the variation in the numbers of SBIR and STTR grant applications over the 2015–24 period is highly correlated with the variation in NIH funding across all grant types (with all correlation coefficients examined being greater than 0.90).
Thus, explicitly taking into account the location of the principal investigator in making funding decisions (as the NIH seems to be considering) wouldn’t necessarily level the playing field. This is because, at least when it comes to the chances of receiving funding as a function of a researcher’s state, the playing field is already fairly flat.
