Our Division of Training, Workforce Development, and Diversity (TWD) supports programs at multiple career stages to foster the development of a strong and diverse biomedical research workforce. This post is the first in a series focused on data from NIGMS training programs and is similar to the ones we have done for our research project grant portfolio. Below, we examine trends in NIGMS applications and awards for the Individual Postdoctoral National Research Service Award (NRSA) (F32) and Pathway to Independence Award (K99) programs. NIGMS also supports institutional postdoctoral awards that include the Institutional Research and Academic Career Development Awards (IRACDA) (K12) and NRSA Institutional Postdoctoral Training Grants (T32) focused in clinical areas, and data on these programs were shared previously.
NIGMS F32 Applications, Success/Award Rates, and Demographic Data – Past 5 Fiscal Years
The F32 fellowship program supports research training of highly promising postdoctoral candidates toward becoming productive, independent investigators in biomedical research fields relevant to the NIH mission. Within the last 5 fiscal years (FY 2015 to FY 2019), NIGMS received approximately 350 to 425 applications each year to the NRSA F32 program (Figure 1).
F32 Applications, Awards, and Success Rates
Figure 1. Number of NIGMS F32 Applications, Awards, and NIGMS and NIH Success Rates, FY 2015-2019. The blue bars (left axis) represent the number of F32 applications, while the orange-striped bars (left axis) represent the number of F32 awards. The solid gray line (right axis) represents the NIGMS success rate, defined as the percentage of reviewed grant applications that receive funding. For comparison to the NIGMS success rate, the NIH success rate is included as a yellow dashed line and indicates that NIGMS had higher success rates over the past 5 years than NIH overall. NIH-wide data are from the NIH Data Book.
Figure 1 shows the NIGMS F32 competing application submissions, awards, and award success rates. A comparison with NIH-wide data indicates that the NIGMS application numbers generally mirror the pattern of NIH-wide F32 applications and the NIGMS success rates are slightly higher than that of NIH overall. Note the decline in awards between FY 2017 and FY 2018 is a result of applications considered at our September 2017 advisory council meeting. These awards were funded ahead of schedule and therefore are reported in an earlier fiscal year.
Figure 2 illustrates yearly award rates by percentile score and indicates that applications at a percentile lower than 25 were most likely to be awarded. Funding decisions are based on multiple criteria and not exclusively on the application’s percentile, as indicated by the fact that some awards are made for applications into the 40th percentile. Additionally, some applicants with favorable review scores (below the 20th percentile) may have accepted other fellowship opportunities and declined the NIGMS F32 award.
F32 Award Rate
Figure 2. NIGMS F32 Award Rate, FY 2015-2019. Each line represents the fiscal year award rate as a function of percentile and shows that awards at percentiles lower than 25 were most likely to be awarded. Applicants who were offered an NIGMS F32 but declined it to accept another fellowship are counted as not having received an award.
Below, we provide the demographic characteristics of NRSA F32 applicants and awardees by race/ethnicity (Figure 3), gender (Figure 4), and age (Figure 5) over the same 5-year period. The numbers of applications and awards interactively displayed are rounded values and are not actual counts. The numbers were systematically rounded up or down to the nearest 10 (for small values) or 25 (for larger values) relative to the actual number. This representation of the data shows overarching trends while ensuring adequate masking of small sample sizes to protect the applicants’ privacy. In comparing award rates, we performed a Fisher’s exact test using funded and unfunded applications across all groups (excluding “Other” or “Not Reported”) to determine whether the differences in award rates across groups are unlikely due to random chance based on simulated alternative distributions. In some cases, we performed follow-up pairwise comparisons to identify potential drivers of observed overall differences.
Figure 3 shows that approximately 80% of applications and awards were among researchers who are white or Asian (well represented). The award/acceptance rates for each racial/ethnic group reveal differences across groups (p = 0.04) with the award/acceptance rate difference between white and Hispanic awardees driving this effect. A larger proportion of Hispanic applicants were among the well-scoring applicants who did not receive/accept an award than applicants from other racial and ethnic groups. Upon closer inspection, most of these well-scoring applicants declined awards in favor of fellowships from other funding organizations, which strongly contributed to the difference in apparent award rates for Hispanic applicants. When combining race/ethnicity data into well-represented and underrepresented groups, we found award rates were higher for applicants from well-represented groups than those in underrepresented groups, at 29% and 22% respectively, primarily driven by the lower acceptance rate for Hispanic applicants, with marginal significance found on statistical testing (p = 0.06). Overall, there appears to be no statistical difference in the awarding rates, but there is a measurable difference in the acceptance rates for Hispanic applicants.
F32 Applications, Awards, and Award Rates by Race/Ethnicity and Representation
Figure 3. Percentages of NIGMS F32 Applications, Awards, and Award Rates by Race/Ethnicity and Representation, FY 2015-2019. The top graph shows the relative percentage of F32 applications and awards by race/ethnic groups. The category “Other” includes individuals who are multiracial or who withheld their race/ethnicity. The second graph from the top shows the award rate by race/ethnic group. Note that the award rates for American Indian or Alaska Native and Pacific Islander groups are not displayed because the numbers fell below the thresholds (n < 11) used at NIH to protect against disclosure in keeping with the Privacy Act of 1974. The third graph from the top shows the relative percentage of F32 applications and awards by well-represented and underrepresented groups. The well-represented group includes white and Asian, and the underrepresented group includes Hispanic, Black or African American, American Indian or Alaska Native, and Pacific Islander. The bottom graph shows the award rate by representation status. Rounded application and award counts can be displayed by hovering your cursor over the graph.
Figure 4 shows that the gender distribution of applications (30% women, 60% men, 10% not reported) is similar to that of the NIGMS early stage investigator Maximizing Investigators’ Research Awards (ESI MIRA) program (27-30% women, 70-73% men). This indicates that this gender gap is also reflected in the applicant pool of the early postdoctoral fellowship awards. The award rates were 30% for men and 28% for women (p = 0.47). This is in keeping with other studies that have shown that women and men have similar success rates in most award programs.
F32 Applications, Awards, and Award Rates by Gender
Figure 4. Percentage of NIGMS F32 Applications, Awards, and Award Rates by Gender, FY 2015-2019. The top graph shows the relative percentage of F32 applications and awards by gender. The bottom graph shows the award rate by gender (p = 0.47). The category “Not Reported” includes individuals who did not provide a gender identity. Rounded application and award counts can be displayed by hovering your cursor over the graph.
Lastly, Figure 5 shows that researchers under age 30 made up almost one-third of applicants and almost 40% of awardees, while researchers between ages 30 and 35 made up almost two-thirds of applicants and awardees, and researchers over 35 made up a small fraction of applicants and awardees. Award rates decreased as age increased, and a comparison of award rates across all three age groups detected statistically significant differences between groups (p < 0.001). As noted in our guidelines for postdoctoral fellowships, NIGMS considers length of time already spent in the sponsor’s laboratory as a factor in making award decisions. Postdoctoral applicants who have been in their sponsor’s laboratory for a long time without a strong justification for the need for this extended period of training are less competitive.
F32 Applications, Awards, and Award Rates by Age
Figure 5. Percentage of NIGMS F32 Applications, Awards, and Award Rates by Age, FY 2015-2019. The top graph shows the relative percentage of F32 applications and awards by age group. The bottom graph shows the award rate by age group. Rounded application and award counts can be displayed by hovering your cursor over the graph.
In addition to the demographic characteristics of NRSA F32 applicants and awardees, we also examined geographic representation, specifically by Institutional Development Award (IDeA) and non-IDeA states. States that are eligible for the IDeA program historically have had low levels of NIH funding, and the IDeA program supports centers and networks to strengthen institutional biomedical research capacity and infrastructure in these states. From FY 2015 to FY 2019, fewer than 2% of NRSA F32 applications were from individuals in IDeA states. The award rate was higher for individuals from non-IDeA than IDeA states, though the difference is not statistically significant (p = 0.07). While the F32 program is not part of IDeA, NIGMS is interested in promoting geographic diversity across its portfolio and encourages eligible applicants from IDeA states to apply to all of the Institute’s programs.
NIGMS K99/R00 Applications, Success/Award Rates, and Demographic Data – Past 5 Fiscal Years
The NIH Pathway to Independence Award (K99/R00) program facilitates timely transitions of outstanding postdoctoral researchers with research and/or clinical doctorate degrees from mentored, postdoctoral research positions to independent, tenure-track or equivalent faculty positions. The program provides support during this transition to help awardees launch competitive, independent research careers.
As is the case across NIH, the NIGMS K99/R00 program is smaller in overall numbers of applications and awards when compared to the NRSA F32 program (compare Figure 1 and Figure 6). The K99/R00 success rate increased from 16% in FY 2015 to 23% in FY 2018, before dropping to 18% in FY 2019. The total number of competing awards has been stable over the last 3 years at 20 awards, and the changes in success rate have been primarily driven by variation in number of applications.
K99 Applications, Awards, and Success Rate
Figure 6. Number of NIGMS K99 Applications, Awards, and NIGMS Success Rate, FY 2015-2019. The blue bar (left axis) represents the number of K99 applications. The orange-striped bar (left axis) represents the number of K99 awards. The gray line (right axis) represents the NIGMS success rate.
Unlike F32 applications, which are reviewed by the NIH Center for Scientific Review, K99 applications are reviewed by NIGMS special emphasis panels and receive only overall impact scores. Applications with overall impact scores lower than 20 were most likely to be awarded (Figure 7). Like the F32 program, funding decisions are based on multiple factors and not exclusively on overall impact score.
K99 Award Rate
Figure 7. NIGMS K99 Award Rate, FY 2015-2019. Each line represents the fiscal year award rate as a function of priority score and shows that applications with an overall impact score lower than 20 were most likely to be awarded.
Demographic trends of the K99 program (Figure 8) indicate that researchers from well-represented groups submitted approximately 80% of applications (approximately 50% from white and 30% from Asian). Men (approximately 60%) submitted a higher percentage than women (about 32%), and the largest share of applications, approximately 75%, came from researchers between ages 30 and 35. Applications from non-IDeA states far outnumbered applications from IDeA states (<1%). Comparisons of K99 award rates across demographic characteristics or geographic representation were not statistically significant, but the small number of applications from underrepresented groups and states makes it difficult to detect possible differences.
K99 Applications, Awards, and Award Rates by Demographic Categories
Figure 8. Percentage of NIGMS K99 Applications/Awards by Race/Ethnicity, Representation, Gender, and Age, FY 2015-2019. The top graph shows the relative percentage of K99 applications by race/ethnic groups. The second graph from the top shows the relative percentage of applications from well-represented and underrepresented groups. The well-represented group includes white and Asian and the underrepresented group includes Hispanic, Black or African American, American Indian or Alaska Native, and Pacific Islander. Award count percentages are not displayed for either of these demographic groupings, as there were fewer than 11 awardees in underrepresented groups. The third graph from the top shows the relative percentage of applications and awards by gender. The bottom graph shows the relative percentage of applications and awards by age. Statistical tests comparing award rates did not reveal significant differences by these demographic characteristics. Rounded application and award counts can be displayed by hovering your cursor over the graph..
Reflections on the F32 and K99/R00 Data
The purpose of this and future data posts is to provide transparency for success rates and demographics of NIGMS’ funded training programs. The demographic data from this post underscore the need to 1.) increase diversity of the F32 and K99/R00 applicant pools, and 2.) take additional steps to ensure equity in the grant application and award decision-making process. We provide a few examples below of how NIGMS is addressing this.
NIH research award rate disparities have been shown to persist over time [PDF] due to a number of factors. One of the initiatives undertaken to address this is the Diversity Program Consortium (DPC), managed by NIGMS. One DPC component is the National Research Mentoring Network, which has developed evidence-informed mentoring, networking, and grant writing workshops for biomedical researchers from diverse backgrounds, such as those from underrepresented groups.
NIGMS supports many initiatives to enhance the diversity of the applicant pool. The Division for Research Capacity Building (DRCB) supports research, faculty development, research training, and research infrastructure improvements in states where levels of NIH research funding have historically been low. NIGMS’ TWD collaborates regularly with DRCB to reach out to individuals conducting research in IDeA states to encourage applications for TWD funding, including the F32 and K99/R00 programs.
NIGMS also supports research to understand and inform interventions that promote the research careers of individuals in biomedical sciences. The research contributes to the evidence base for effective, high-impact, scalable interventions that improve the success of individuals from diverse backgrounds pursuing independent biomedical research careers.
The data presented here confirm a trend showing that women’s overall representation in academic biomedical sciences declines between training and independent scientist career stages, including during the postdoctoral research phase, which is then reflected in their lower representation among NIH research grant applicants and awardees compared to men. In addition to our diversity-enhancing programs and supporting research to understand and inform interventions that increase persistence in the biomedical research workforce, NIGMS contributes to the NIH Working Group on Women in Biomedical Careers, which strives to understand the barriers women face in career advancement in biomedical research and identify possible approaches that NIH can take to help overcome those barriers.
NIGMS has a long tradition of supporting diversity-enhancing programs from the community college level to the postdoctoral level; however, we recognize that both NIGMS and our institutional partners can do more to promote diversity in the biomedical workforce, particularly at critical junctures in researchers’ careers such as the postdoctoral and early career phase. NIGMS wants to work with institutions to achieve this goal and has already taken some steps to make additional progress. The mismatch between the increasing proportion of underrepresented biomedical Ph.D. earners and comparable representation among biomedical faculty was one reason that NIGMS recently launched the Maximizing Opportunities for Scientific and Academic Independent Careers (MOSAIC) program. This program aims to facilitate the timely transition of promising researchers from diverse backgrounds from their mentored, postdoctoral research positions to independent, tenure-track or equivalent faculty positions at research-intensive institutions. To date, the MOSAIC program has already received more applications from Black and African American applicants than the total number received in the last 5 years of the NIGMS K99 program. By actively and regularly assessing our training portfolio, NIGMS will ensure that programs such as MOSAIC meet their objectives and make inroads to address this long-standing challenge in the biomedical workforce.
In addition to studying the demographics of those awarded each type of training award, it would also be helpful to observe the demographics of the host lab (the postdoc phase in the case of K99s).