K99-R00 program: F32 awardees as a control group

Nov 06 2014 Published by under Uncategorized

I recently posted a series of analyses on the K99-R00 program, particularly the FY2007 cohort (the first year of the program). In my initial post on this program, I noted that recipients of F32 awards during the same period might make a reasonable group to compare to the K99-R00 recipients. I have now performed some of this analysis.

In FY2007, there were 612 new F32 awardees across NIH. Of these, 143 have some post-F32 award in NIH RePORTER. However, these investigators do not reflect the full F32 cohort who have gone into academia as others may be in academic positions, but have not received NIH grants. Unfortunately, information for these individuals is not readily available through a public database. However, through the wonders of the internet, search engines, and social media sites such as LinkedIn, I was to able to identify the likely current positions of approximately 90% of the FY2007 awardees. The representation of those in academic positions is likely close to 100% given the higher web profiles of those in academia compared with those in other sectors.

I was able to identify 304 individuals with an academic positions likely suitable for applying for grants. Thus, it appears that approximately 50% of the F32 awardees transitioned to academic positions after the completion of their training. This compares with 93% of the K99 awardees who transitioned to the R00 phase as well as several others who moved to academic positions (often outside the United States) without activating the R00 phase of the K99-R00 award. Thus, it appears that while the K99 mechanism was nearly 100% efficient in transitioning investigators to academic positions, only about half of this F32 cohort made this transition.

Other individuals have moved to positions in the pharmaceutical and biotech industries, some physicians have moved into private practice settings, and some  individuals have scientific staff positions. Approximately 25 individuals still appear to be in listed in positions categorized as postdocs, even 7 years after the initial F32 award.

Did the F32 cohort come from the same distribution of institutions as the K99 cohort? The distribution of institutions, based on rankings of NIH funding levels, are shown below:

F32-K99 institution plot

These distributions are relatively similar to one another with the median institutional rank for the F32 awardees at 33 compared with 23 for this cohort of K99 awardees. Given the imperfections of institutional NIH funding rank as a measure as well as the relatively small sample sizes, it is not clear how significant this difference is.

How about the distribution of the academic positions obtained by the F32 awardees compared to the K99 cohort? These are compared below:

F32-K99 Academic Institutions

These distributions are more different. In particular, the F32 awardee distribution shows a significant number of investigators who work at institutions with relatively low NIH funding rank. These are largely institutions that have quite substantial undergraduate teaching missions.

What fraction of the F32 cohort have obtained R-mechanism or similar funding? 86 investigators have obtained at least one R01, DP2, R21, R15, R03, or R56 award and two have obtained positions and funding within the NIH intramural program. Of these, 60 of these have received R01 or DP2 (NIH Director's New Innovator) awards.

In addition, a considerable number of this cohort received K awards subsequent to their F32 award including 33 K99 awards and 41 other (K01, K07, K08, K22, K23) K awards.

For comparison, 120 of the 182 2007 K99 cohort members have obtained at least one R01, DP1, DP2, R21, R15, R03, or R56 award and one moved to the NIH intramural program. Of these, 108 of these have received R01, DP1, or DP2 awards.

The overall funding patterns as a function of time are compared below:

F32 K99 funding curves

These curves show the overall amount of funding in different categories (F32, K awards except K99, K99, R00, R01 equivalent (R01, DP1, DP2), and all other awards) normalized to the number of investigators in each cohort. Note that the scale is 5 times larger for the K99 cohort than for the F32 cohort. These curves show the initial investment in F32 or K99 awards, respectively, followed by the time evolution of other types of funding.

The amount of R01 equivalent funding per investigator for the K99 cohort is nearly six times higher than that for the F99 cohort. This is a consequence of the two-fold difference in obtaining academic positions (~100% versus ~50%) and the three-fold (108/182 versus 60/304) difference in obtaining R01 given an academic position.

It also appears that the K99-R00 cohort are become independent investigators earlier, certainly as measured by the onset of R00 funding and also as measured by a one- to two-year shift in the time for obtaining R01 or other funding.

Update:  A discussion on Twitter involved examining the costs and outputs for the F32 and K99 programs.

Here are the data related to the graphs above:

F32 program:  Total cost of F32s: $64.89M; K Awards: $27.85M; R00: $20.88M; R01 equiv: $67.61M; Other grants, $29.91M

K99 program: Total cost of K99s: $28.86M; R00: $125.68M; R01 equiv: $149.02M; Other grants, $30.65M

If one considers the R00 total as a cost (since this is a commitment made by NIH at the time of the K99 award), the ratio of Output (Ro1 equiv + Other grants) to Cost (F32 + K + R00) is (67.61+29.91)/(64.89+27.85+20.88) = 0.86 for the F32 program. Recall that 33 F32 recipients in this cohort went on to receive K99 awards.

Similarly, Output/Cost = (149.02+30.65)/(28.86 + 125.68) = 1.16

Of course, R00 represents research grants, so they could also be considered as output rather than cost. Under these conditions, Output/Cost = (67.61+29.91+20.88)/(64.89+27.85) = 1.28 for the F32 program and Output/Cost = (149.02+30.65+125.68)/(28.86) = 10.58.

As a compromise between these views, the R00 value could not be considered as either a cost or an output. Under these conditions, Output/Cost = 1.05 for the F32 program and 6.23 for the K99 program.

Note that these Output/Cost ratios are not fixed, but will increase over time as outputs grow while Costs are fixed.

12 responses so far

Leave a Reply