Cardiothoracic and Vascular Surgeons Achieve High Rates Of K-Award Conversion Into R01 Funding

Abstract

Background:

Obtaining National Institutes of Health (NIH) R01 funding remains extremely difficult. The utility of career development grants (K awards) for achieving the goal of R01 funding remains debated, particularly for surgeon-scientists. We examined the success rate for cardiothoracic and vascular (CTV) surgeons compared to other specialties in converting K-level grants into R01 equivalents.

Methods:

All K (K08 and K23) grants awarded to surgeons by the NIH between 1992-2017 were identified through NIH RePORTER, an online database combining funding, publications, and patents. Only grants awarded to CTV surgeons were included. Grants active within the past year were excluded. Mann-Whitney U-tests and Chi-squared tests were used to compare groups.

Results:

A total of 62 K grants awarded to CTV surgeons were identified during this period. Sixteen grants were still active within the last year and excluded from analysis. Twenty-two (48%) of the remaining K awardees successfully transitioned to an R01 or equivalent grant. Awardees with successful conversion published 9 publications per K grant compared to 4 publications for those who did not convert successfully (p=0.01). The median time for successful conversion to an R grant was 5.0 years after the K award start date. Importantly, the 10-year conversion rate to R01 was equal for CTV surgeons compared to other clinician-investigators (52.6% vs 42.5%).

Conclusions:

CTV surgeons have an equal 10-year conversion rate to first R01 award compared to other clinicians. These data suggest that NIH achieves a good return on investment when funding CTV surgeon-scientists with K-level funding.

Introduction

NIH funding has consistently become harder to obtain. Surgeon-scientists face an uphill battle in obtaining NIH funding due to their clinical workload and need to operate in order to maintain their surgical skills. Additionally, surgeons have very little dedicated time to perform research throughout their training (1). The NIH has established Mentored Clinical Scientist Research Career Development Awards (K08s) and Mentored Patient-Oriented Research Career Development Awards (K23) for physicians to develop skills for successful scientific research. These awards generally last 5 years and awardees are required to work with a senior scientific mentor in order to develop investigative skills to obtain independent NIH funding. In the traditional academic environment, R01 grant funding is typically required to maintain a successful research lab. Most junior surgical faculty have limited dedicated research time during residency and most have no formal curriculum (2). In a study at a the University of Toronto, Goldenberg and collogues concluded that significantly fewer current day surgery residents are pursuing research compared to previous years (3). If this trend continues, the number of surgeons who pursue basic science research will continue to decline.

From 2007-2010, only 16% of K08s were awarded to faculty in Departments of Surgery, which is indicative of the current funding environment (4). To combat low funding and establish increased mentorship, the Thoracic Surgery Foundation for Research and Education (TSFRE) (2002) and the Society for Vascular Surgery (SVS) Foundation (1999) have established joint programs with the National Heart, Lung, and Blood Institute to fund career development programs to thoracic, cardiac, and vascular surgeons (5, 6).

Given societal support, we hypothesized that CTV surgeon-scientists would achieve a high level of success in converting K-level awards (K08 and K23) to R01 or equivalent grants even in an era of difficult NIH funding. We further hypothesized that CTV surgeons have overcome significant clinical demands to establish successful research laboratories. To this end, we determined the conversion rate of K to R funding for CTV surgeons compared to published benchmarks. We further evaluated the characteristics of successful K converters compared to unsuccessful converters to establish a baseline for budding surgeon-scientists interested in pursuing translational science research.

Methods

Collection of Grant Information

The National Institutes of Health Research Portfolio Online Reporting Tools (NIH RePORTER) was queried to collect grant information from 1992 to present. NIH RePORTER is a database containing research project information, funding records, project description, citations of papers supported by the grant, and patent information from the U.S. Patent and Trademark Office. The “Project Details” search option was used to search for all K grants awarded to surgery departments. Only K08 and K23 awards were included in our analysis due to the basic science focus of these awards. A Python script containing Phantom JS and Beautiful Soup packages was used to capture the following information about each grant: identifier number, title, grant type, principal investigator, awardee organization, awardee department, grant start/end dates, awarding institute, awarding study section, total funding amount, number of publications, PMIDs of publications, and the journal of each publication. The length of time for transition from K award to R01 award or equivalent was calculated as the time between the K award project start date and the R01 award project start date.

Collection of Awardee Information

The following awardee information was collected: gender, academic rank, academic department and division, clinical specialty (Cardiac, Thoracic, or Vascular), degrees, institution, and other NIH grant funding. The majority of information was obtained from CTSNet and university faculty pages of surgeons. If an awardee transitioned to private practice, publicly available information online was used to collect awardee information.

Calculation of Grant Impact Metric

To assign impact factor for each publication, Journal of Citation Reports (2016 Journal Citation Reports® (Clarivate Analytics, 2017) Philadelphia, PA) was utilized (7). To accurately weight the amount of effort in each publication, an adjusted Grant Impact metric was calculated. An adjusted Grant Impact takes the impact factor of each paper and weights it according to the author contribution. A first author or last author publication remains the same (1.0 x impact factor), a second or second to last author paper receives a slightly lower score (0.5 x impact factor), and any other paper receives a low score (0.2 x impact factor). Schroen et al. have used this method for evaluating authors’ actual impact via publications (8). A Total Grant Impact is the sum of all adjusted Grant Impact metrics from the papers in a particular grant. Grant Impact Per Year was calculated by dividing the Total Grant Impact by the length of funding for each grant. Grant Impact Per Dollar was calculated by dividing the Total Grant Impact by the total funding of each grant (per $100,000).

Statistics

Categorical variables are reported as n (%) and continuous variables were reported as median [IQR]. Groups were compared using a Mann-Whitney U Test for continuous variables and Chi-Squared test for categorical variables as appropriate. Statistical tests and figures were performed using GraphPad Prism Version 7.00 for Macintosh (Copyright © GraphPad Software, La Jolla California, USA).

Results

K-Awardee Characteristics

A total of 62 K Award (K08 and K23) recipients were identified whose primary specialty lies in cardiac, thoracic, or vascular surgery (Figure 1). The 62 K grants were awarded by the following institutes: National Heart Lung and Blood Institute (NHLBI: 52); National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK: 5); National Cancer Institute (NCI: 3); National Institute on Aging (NIA: 1); and Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD: 1). Of these 62 recipients, 22 awardees successfully converted to an R01 (21) or equivalent P50 (1; $2.5 million) and 40 awardees have not obtained R01-level funding to date. In the successful conversion group, one awardee was excluded due to academic fraud. Within the 40 K awardees who had not transitioned to R01-level funding, 16 K awards are still active or were active within the last 12 months of the present analysis and were excluded. Of the remaining 24 grants, 17 K awardees had not successfully obtained R01-level funding after 10 years from K award start date, while 7 grants have not yet reached the 10-year threshold after the K award start date.

Figure 1:

CONSORT style diagram of K award recipients. A total of 62 CTV K awards were identified with 22 obtaining successful R01 award (or equivalent) conversion. Of the 62 awards, 40 did not obtain R01 award (or equivalent) conversion. Of the 40 with no successful conversion, 16 awards were active within the last 12 months and were therefore excluded to leave 24 K awardees with no R01 award conversion. One award was excluded from the R01 award conversion group due to academic misconduct, which led to a successful R01 award conversion group of 21 with all 21 candidates obtaining R01-level funding within 10 years. Of the 24 unsuccessful K awardees, 7 awardees have not crossed the 10-year mark from the start of their K award but 17 have surpassed that threshold. Six of the 7 awardees that have not passed the 10-year mark are now in private practice compared to 1 of the 17 awardees that have passed the 10-year mark.

There was no significant impact of gender between groups of awardees successfully proceeding to R01-level funding compared to awardees not achieveing R01-level award conversion (Table 1). We found no significant differences in the number of Thoracic, Cardiac, or Vascular surgeons affecting the rate of K award to R01 award conversion (Thoracic: 2 (9.5%) vs. 1 (4.2%); p = 0.472, Cardiac: 6 (28.6%) vs. 5 (20.8%); p = 0.547, and Vascular: 13 (61.9%) vs. 18 (75.0%); p = 0.344). Successful K award converters had previously obtained 6 (28.6%) F32 NIH Ruth L. Kirschstein National Research Service Award Individual Postdoctoral awards while in their training prior to receiving their K award compared to 1 (4.2%) member of the unsuccessful K award converter group (p = 0.028). Successful K-converters have a higher number of additional degree holders (7 (33.3%) vs. 4 (16.7%); p = 0.194) and a higher number of dual-degree MD/PhD investigators (6 (85.7%) vs. 2 (50.0%); p = 0.077). Next, we evaluated the number of awardees obtaining leadership positions in each group during their career. Successful K to R01 award converter principal investigators obtained 9 (42.7%) leadership positions compared to 12 (50.0%) leadership positions of their counterparts (p = 0.632). Finally, we evaluated the current practices of K awardees and found that no one who successfully converted their K to an R01 (or equivalent) currently practice in a private setting, while 7 (29.2%) of the unsuccessful K to R01 awardees are currently in a private practice setting (p = 0.007).

Table 1:

K-Awardee Characteristics

Characteristic	Successful K to R01 Award PIs (n = 21)	Unsuccessful K to R01 Award PIs (n = 24)	p-value
Awardees	21	24	-
Male	18 (85.7%)	22 (91.7%)	0.526
Female	3 (14.3%)	2 (8.3%)	0.526
Thoracic Surgeons	2 (9.5%)	1 (4.2%)	0.472
Cardiac Surgeons	6 (28.6%)	5 (20.83%)	0.547
Vascular Surgeons	13 (61.9%)	18 (75.0%)	0.344
F32 Awards Obtained Before K Award	6 (28.6%)	1 (4.2%)	0.028
Additional Degree Holders	7 (33.3%)	4 (16.7%)	0.194
PhD	6 (85.7%)	2 (50.0%)	0.077
M.S.	1 (14.3%)	2 (50.0%)	0.603
MBA	1 (14.3%)	0 (0%)	0.280
Leadership Position a	9 (42.7%)	12 (50.0%)	0.632
Vice Dean	1 (11.1%)	0 (0%)	0.280
Department Chair	3 (33.3%)	1 (8.3%)	0.234
Department Vice- Chair	1 (11.1%)	1 (8.3%)	0.923
Division or Section Chief	4 (44.4%)	10 (83.3%)	0.102
Currently Private Practice	0 (0%)	7 (29.2%)	0.007

Chi-squared test was used for p-value calculation.

^a)Leadership position held at any point.

K-Award Characteristics

We analyzed the characteristics of each K award (Successful Conversions: 20 K08s and 1 K23, Unsuccessful Conversion: 20 K08s and 4 K23s). There was no significant difference in length of funding (4.9 vs. 5.0 years; p = 0.32) or total funding for each grant ($628k vs. $615k; p = 0.30) between the two groups. However, successful K awardees published more papers (9 vs. 4; p = 0.01). Successful K awardee converters obtained significantly higher Total Publication Impact per grant (34.46 vs 11.79; p = 0.0001), Impact Per Year (5.75 vs 2.17; p = 0.0002), Impact Per $100,000 (5.31 vs 2.00; p = 0.002), and Impact Per Publication (3.70 vs 3.13; p = 0.025).

Subsequent Grant Characteristics

K awardees who successfully obtained R01-level funding following their K award obtained 42 grants following their award compared to 2 grants for the awardees who did not obtain R01-level funding (p < 0.0001). Thirty of the 42 obtained grants were R01 grants (0 for Unsuccessful K awardee group p < 0.0001). Other less commonly obtained grants included R56 (2), I01 (3), R21 (1), R43 (1), R44 (1), P30 (1), P50 (1), UM1 (1), and U01 (1) grants. One awardee in the unsuccessful conversion group obtained 1 R56 grant and 1 R03 grant. Finally, successful K awardees obtained significantly higher funding following their K award compared to their counterparts ($82 million vs. $558 thousand; p < 0.0001).

Timeline of Achieving Funding

Finally, we evaluated the length of time between obtaining a K award and successfully converting to an R01.. When considering all K awardees, each successful K to R01 converter obtained their R01 award within 10 years of beginning their K award (Figure 2). The median time to achieve R01-level funding for successful K award converters was 5.00 years

Figure 2:

Kaplan-Meyer curve of the time to conversion of K awardees (n = 38). Unsuccessful K awardees who have not crossed the 10-year mark of obtaining their K award are not included. The median time to R01 award conversion for successful K to R01 converters is 5.00 years.

Comment

This study has determined the conversion rate from NIH Mentored Career Development Awards (K Awards) to NIH Research Project Grant Program awards (R01 Awards) for CTV surgeons. We have shown that the 10-year conversion rate to R01 or equivalent (P50) awards for CTV surgeons is 55.3 % (21/38). When considering strict K award to R01 grant conversion, CTV surgeons have a 10-year conversion rate of 52.6 % (20/38). This compares favorably to a previously published K to R01 conversion rate of 42.5% in 2,784 investigators by Jagsi et al. (9).

After determining that CTV surgeons obtain a high rate of R01-level funding after receiving a K award, we sought to characterize successful K awardees and unsuccessful K awardees (Table 1). Interestingly, there were no differences in gender distribution, specialty of CTV surgeon, and leadership positions held. We evaluated how many awardees had received an F32 (NIH Ruth L. Kirschstein National Research Service Award Individual Postdoctoral Fellowship) prior to their K award and found that 6 (28.6%) successful K awardees had received an F32 compared to 1 (4.2%) unsuccessful K awardee (p = 0.028). A structured research program with emphasis on mentorship and training environment is required for F32 funding. These trainees most likely benefitted from early rigorous research training which may have conferred an advantage when applying for and receiving both K and R awards. Next, we anticipated there would be a greater number of leadership positions held by surgeons who had R01-level funding due to research productivity being an important component for promotion based on prior studies, but there was no association with successful conversion (10). Unsuccessful K to R01 converters may obtain more leadership positions by choosing to focus on other merits required for promotion such as teaching residents/medical students, clinical research, advanced clinical positions, and mentorship of younger faculty. Additionally, faculty already in leadership positions may not pursue K to R conversion. However, there was a slight difference in the number of successful K awardees who have a PhD. This was surprising as 43% of all K award recipients had received a PhD, but amongst CTV surgeons, the K to R award conversion group seems to have a slightly higher cohort of PhD holders ((28.6%) vs (8.3%); p = 0.077) (4). MD/PhD surgeons receive extensive basic science training during their PhD and this training likely contributes to their success in not only obtaining K grants but also converting to R level awards. Finally, we evaluated the number of awardees from each group who transitioned from academia to private practice. No surgeons who were successful in obtaining R01-level funding transitioned to private practice, but 7 of the 24 unsuccessful K awardees moved to private practice (p = 0.007).

While there were a few differences in the characteristics of the awardees themselves, the grants in each group had notable differences. Successful K grant conversions had a significantly higher number of publications and publication impact (total, per year, per $100,000, and per publication). While we expected successful grant conversion to perform better, we did not expect a clear delineation between successful and unsuccessful conversion. Successful K award converters on average published 5 more papers with a Total Publication Impact of 22 points higher compared to unsuccessful K award converters. We noticed that not only are successful K awardees publishing more papers, but these publications are generally in higher impact journals compared to their counter parts. From this data, we conclude that successful K award conversion requires not only a high quantity of papers, but also high quality papers.

Next, we evaluated the subsequent grants obtained by all K awardees. Successful K awardees obtained 42 grants including 30 R01 grants (p < 0.0001). Successful K awardees obtained over $81 million in their subsequent 42 NIH grants. While unsuccessful K award converters did not obtain R01-level funding, we did not expect them to obtain very little NIH funding ($558 thousand; p < 0.0001). One possibility for this disparity is that K awards are intended to be a conduit for obtaining R01-level funding and principal investigators do not apply for or successfully compete for non-R01-level grants if they do not initially get R01-level grants.

The median time to conversion for successful K awardees converters was 5.00 years, approximately the same length as the K award itself (Figure 2). Interestingly, not a single successful conversion occurred beyond 10 years after initiating a K award. This inability to convert from a K to an R award is likely due to surgeon-scientists phasing out of translational research as they run out of funding. In a publication regarding surgeons obtaining funding, Kron has emphasized “most successful investigators have been turned down on multiple grants, have continued to do their research, and have reapplied for funding until successful” (11). One of the largest barriers to surgeons obtaining K award funding is a lack of applications. Surgeons do not submit many K08 grants (7.1% of applications 1990-2005) and are last in K23 grants applications (2.2% of applications 2000-2005) (12). In 2016, the success rates for all K08 applications were 31.3% for NCI and 44.3% for NHLBI. In 2016, the success rates for all K23 applications were 16.1% for NCI and 44.6% for NHLBI (13).

Mentorship is one of the most critical aspects of K awards. K awardees are required to have an established mentor in their field of study to guide them through the process of establishing a strong research enterprise. Michael Pasque states that mentees should have goals such as “literature vetting of research hypothesis, early procurement of startup funding, acquisition of preliminary data, and formal grant submission”(14). Mentorship is critical in the development of a surgeon’s operating skills and we argue that K award mentorship is critical to the development of a surgeon-scientist (15). Young academic surgeons need dedicated research time to lead projects and write grants (16). Cardiothoracic surgeons receive very low per capita NIH funding. Ratcliff et al. have argued that this is due to few applications per capita and have advocated for junior faculty applying for career development awards (17). We have delineated the factors of what separates a successful surgeon-scientist with a career development award (K award) from an unsuccessful one in terms of productivity during the K award years. Mentorship during the K award period is associated with increased productivity and therefore junior faculty interested in leading an NIH funded group should aim to find mentors quickly to develop their basic science research laboratory (18).

Important limitations of this study include the small sample size of our two groups. This limitation is further amplified by the fact that there are far fewer applications and successfully funded K awards to surgeons. One of the aims of our study was to highlight the success of CTV surgeons and to advocate for greater K award funding. A second limitation of our study is that we were not able to survey K awardees for their personal experiences with their approach to basic and translational science during their junior faculty years. A survey would have elucidated mentorship, research atmosphere, protected research time, and goals at the time of their K award. There may have been special circumstances in which unsuccessful K awardees did not attempt to obtain R01-level funding due to clinical or personal constraints. We were not able to obtain data on how many applications K awardees submitted in order to obtain R01-level funding or how NIH scored those grants. Finally, we were not able to obtain a success rate of all R01 grant applications for CTV surgeons or the total number of R01s obtained by CTV surgeons to establish a baseline for this study.

In conclusion, we have demonstrated that cardiothoracic and vascular surgeons achieve a high rate of conversion from K-level career development awards to independent R01-level awards comparable to all other peers. We have shown that successful K awardees that convert to R01 awards produce highly impactful publications and are extremely productive during their 5-year K award in order to obtain further NIH funding. K awardees that are not successful in obtaining R01-level funding failed to obtain any significant NIH funding. Finally, we highlight that the first decade is the most important as all successful K awardees obtained R01-level funded prior to reaching the 10-year mark from the start of their K award. K awardees in surgery should aim to publish impactful papers early in the lifespan of their K award so that they are competitive during the R01 award application process. In a time of declining NIH funding, CTV surgeons are proving to be prolific researchers who should be increasingly funded by the NIH.

Table 2:

K Grant Characteristics

Each Grant	Successful K Conversion Grants (n = 21)	Unsuccessful K Conversion Grants (n = 24)	p-value
Length of Funding (years)	4.918 (4.81 – 5.00)	5.00 (4.87 – 5.94)	0.315
Total Funding	$627,890 ($606,431 - $654,615)	$614,925 ($532,035 - $644,694)	0.304
Publications	9 (5 – 19)	4 (2 – 8)	0.012
Total Publication Impact	34.46 (21.41 – 72.38)	11.79 (6.67 – 23.83)	0.0001
Per Year	5.75 (4.41 – 15.29)	2.17 (1.35 – 4.92)	0.0002
Per $100,000	5.31 (3.30 – 10.66)	2.00 (1.27 – 4.22)	0.002
Per Publication	3.70 (2.67 – 4.88)	3.13 (1.78 – 3.75)	0.025

Median (Interquartile Range)

Mann-Whitney U test was used for p-value calculation.

Table 3:

Subsequent Grant Characteristics

Characteristic	Successful K to R01 Award PIs	Unsuccessful K to R01 Award PIs	p-value
K Awards	21	24	-
Total Grants a	42	2	<0.0001
R01b	30 (71.4%)	0 (0%)	<0.0001
R01 Competitive Renewals	4 (9.5%)	0 (0%)	0.040
I01c	3 (7.1%)	0 (0%)	0.094
R56	2 (4.8%)	1 (50.0%)	0.592
R03	0 (0.0%)	1 (50.0%)	>0.999
R21	1 (2.4%)	0 (0%)	0.467
R43	1 (2.4%)	0 (0%)	0.467
R44	1 (2.4%)	0 (0%)	0.467
P30	1 (2.4%)	0 (0%)	0.467
P50	1 (2.4%)	0 (0%)	0.467
UM1	1 (2.4%)	0 (0%)	0.467
U01	1 (2.4%)	0 (0%)	0.467
Total Grant Funding b	$81,553,992	$557,600	<0.0001

Total values shown. P-value calculated using Mann-Whitney U-test

^a)T32, KL2, TL1, UL1, M01 grants were not included. Grants obtained before the K award start date were not included.

^b)R01 competitive renewals were counted as a separate grant.

^c)Funding amounts not available for I01 grants funded for Veterans Hospital Projects.