Even after passing through the bottleneck between a postdoc and a faculty position, early career researchers face an uncertain future. This “survival curve” (found on slide 47 in this deck from Mike Lauer) shows the retention rate of NIH PIs since their first award (more RPG survival curves are shown in a 2014 blog post by Sally Rocky). The rate of dropping out is particularly acute 4-5 years after a PI’s first award. In addition, fewer investigators who entered the system between 2001 and 2005 remained at this stage compared to those in a cohort 5 years ahead of them.
As the data above suggest, the 4-5 years after an investigator’s first award is a particularly vulnerable period – possibly becoming even more so over time. Therefore, it’s not surprising that many commenters have voiced support for the GSI on the grounds that it might improve career stability for early and mid stage investigators, a goal that the NIH has sought to implement in the plan’s successor, the NGRI. Unfortunately, this substitution leaves what were, in my view, two of the most compelling reasons for a policy intervention out in the cold: the present inefficiency in use of taxpayer dollars (ie, diminishing returns), and the human capital cost of an unsustainable biomedical workforce structure. I’ll expand a bit on each.
The GSI cap was based on a corpus of literature showing that, as concurrent research funding to a particular group increases, diminishing returns are reaped in terms of numbers of publications, their impact factor, etc. This literature is reviewed in the slide deck from Mike Lauer mentioned above (see slide 34). Diminishing returns are also seen in this analysis of lab productivity in the MIT biology department. While productivity in general increases with more funding, the evidence suggests that the NIH would be able to support more scientific output by spreading the funds more equitably among its grantees. As scientists, our primary concern should be the advancement of science as a whole – not the power that can be amassed in an individual career. Some argue that research produced in large labs is more “transformative,” I’d like to see evidence that this is true on a per-dollar basis.
Unsustainable workforce structure (and poor outcomes for trainees)
Not only are large labs on average inefficient in terms of producing papers, they are also not making good use of human capital: they are producing independent NIH-funded investigator at the same rate as smaller labs per mentor, despite presumably having many more postdocs. (The disturbing image below is slide 37 in slide deck from Mike Lauer). Moreover, these labs in general do not represent a sustainable workforce structure. Alberts, Kirschner, Tilghman, and Varmus have called for a reduction in lab size to bring our research enterprise into equilibrium. Since students and postdocs are supported mainly off of R grants, the most efficient way to reduce lab size would be to cap the funds concurrently available to an individual investigator.
The NIH began to respond to very valid concerns about how training and collaboration were incentivized (or not) with the GSI, but unfortunately, under pressure from well-funded investigators (to get a flavor of disturbing tenor of the objections, see this Boston Globe article), they scuttled the GSI in favor of the NGRI.
The NGRI will cost an estimated $1.1 billion per year, but it is unclear how this policy can be implemented or enforced across all ICs. By contrast, the GSI cap was a clear policy that not only supported early career investigators but also addressed overall research efficiency and promoted sustainable lab sizes. I hope that focus on these urgent problems is not lost.
The views and opinions expressed in this blog are the views of the author(s) and do not represent the official policy or position of ASCB.