|Home | About | Journals | Submit | Contact Us | Français|
Mentors play important roles in training new investigators. This study was designed to determine characteristics of NIH mentored K award recipients and their mentors, their interpersonal interactions, and the factors, which influence satisfaction within this relationship.
A survey of 3027 NIH mentored K recipients and 1384 mentors was conducted in 2009. Nine hundred twenty-nine (30.7%) of the K recipients and 448 (32.4%) mentors completed the survey.
The gender of K respondents was evenly divided while the mentors were 72.1% male. The overall rating of their mentors was positive. Ideally, both thought the mentor should be important in research training; however, in actual practice, both rated the importance as lower. A total of 88.2% of recipients were satisfied with their relationship. Although the number of black K recipients was low, this group was more likely to be dissatisfied with the mentor relationship (6/29 or 20.7%) than their white counterparts. The frequency of meeting or communicating was correlated with K recipient satisfaction.
Overall K recipients are satisfied with their mentor relationships. Although the number of black K recipient respondents was small, the higher level of mentor dissatisfaction should be further evaluated. Qualities of mentors, including the frequency of interactions and accessibility, can influence satisfaction.
Mentors have been described with varying accolades and lofty goals. Daloz said, “Mentors are guides. They lead us along the journey of our lives. We trust them because they have been there before. They embody our hopes, cast light on the way ahead, interpret arcane signs, warn us of lurking dangers and point out unexpected delights along the way.”1 Mentors develop formal and informal relationships with those who request or require guidance. In academic settings, the role of the mentor in ensuring professional transitions and success is well documented.2–6 The literature provides mentor characteristics and attributes that are thought to contribute to effective academic mentoring.7 The role of the research mentor is gaining increased attention.8
The NIH document, “A Guide to Training and Mentoring in the Intramural Research Program at NIH,” defines a research mentor as “a person who has achieved career success and counsels and guides another for the purpose of helping him or her achieve like success.”9 In a review of research mentoring outcomes, 8 the positive outcomes include publications, 10,11 increased time conducting research,10 increased grants,10 higher self-rated research skills,4 academic advancement,12 and career satisfaction. 4 However, there is little information on the characteristics of successful research mentoring relationships. Using a sample of NIH mentored investigators this study explores that relationship.
Through the Kseries of grants, the NIH provides new investigators with salary support for 3–5 years for protected research and training time.13 Four of these K awards mandate an assigned mentor: K25 Mentored Quantitative Research Career Development Award; K23 Mentored Patient Oriented Research Career Development Award; K08 Mentored Clinical Scientists Development Award; K01 Mentored Research Scientist Development Award. Each of these K applications requires designation of a primary mentor who is responsible with the applicant for the planning, direction, and execution of the trainee’s research and training program. While the K 25 guidelines state that a mentor is only required to assist with these areas, the K23, K08, and K01 guidelines precisely define who may serve as the proposed mentor. The mentor should be recognized as an accomplished investigator in the proposed research area and have a track record of successes in training independent investigators in research. The mentor should have sufficient independent research support to cover the costs of the proposed research project in excess of the allowable costs for this award.13
Mentors, therefore, are not only sharing knowledge but are also sharing their research resources. There is no salary or other support for the mentor provided by these K awards, yet the expectations for providing resources and guidance on the science and ethical practice of research to the K recipient are high.
The K recipients represent a cohort of future researchers who have successfully applied for NIH funding. Therefore, they provide a unique opportunity to describe the research mentor relationship. In this study we sought to examine the demographic characteristics of K recipients and mentors, the attributes of the mentoring relationship, how satisfied the K recipients were with their mentors and what factors had an impact on this satisfaction. We expected our results to provide new information that can be used to better understand the mentor–mentee relationship mandated by the K award mechanism. This information may serve to better inform decisions and policy regarding the unique mandate of mentored career awards.
This national study was approved by the Institutional Review Board at Virginia Commonwealth University. As part of a larger study of K01, K08, K25, and K23 award recipients (K’s) and their mentors, 3027 K recipients who were actively funded for their K award in 2007 were identified and their email addresses obtained using the publicly available CRISP database (Computer Retrieval of Information on Scientific Research). The list of K recipients was compiled in 2008 to assure that they had completed at least 1 year of their K prior to our survey. Because there was no way to directly ascertain mentors of K awardees, a two-step identification process was required. A brief Web-based survey was first distributed to the K recipients via email asking for their mentors’ contact information. A total of 1384 (45.7%) responded with information on their mentor. The full survey was then distributed in 2009 to all identified K recipients (N = 3027) and the mentors for whom contact information had been received.
The approach developed by Dillman was modified with follow up emails 1 and 2 weeks after the initial email to those who had not responded.14 Because of the NIH notification of American Recovery and Reinvestment Act (ARRA) challenge grants at the same time as the initial survey email invitation and notification by several potential participants that they would participate after those deadlines, the reminder emails were delayed to allow the final reminder to arrive following the main ARRA challenge grant deadlines.
Nine hundred twenty-nine (929) K recipients (30.7% of the 3027 identified K’s) completed the survey. Four hundred and forty-eight mentors (32.4% of identified mentors) responded to the survey representing 14.8% of the overall mentors for the 3027 K’s. It is acknowledged that these early career investigators may have more than one mentor, however only the primary mentor for the K recipient was surveyed. Because mentors may also have several K recipients, for consistency, they were asked to respond to the questions regarding their most senior K recipient.
Respondents were asked to complete a survey conducted using the Inquisite8 software package (Allegiance, Austin, TX, USA). This survey included questions on demographics; the frequency and content of communication with their mentor; and the mentor’s importance, both ideally and actually for skill training. The John Hopkins’ Mentorship Profile Questionnaire, which looks at outcomes of the mentoring, and the Mentorship Effectiveness Scale, which looks at the qualities of the mentoring, were also included in the survey.2 These scales assess general mentoring qualities, the nature of the mentoring relationship, and outcomes such as publications, grant writing, and conducting research. The Mentorship Effectiveness Scale developed by Berk et al. was used in its entirety, although the rating options were collapsed from six categories that ranged from strongly disagree to strongly agree to four responses (strongly agree, agree, disagree, strongly disagree). Doing so eliminated the slightly disagree and slightly agree responses to be more consistent with the rating scales of other questions in our survey. The surveys were tested with local K recipients and postgraduates and mentors using a think aloud method prior to deploying the national survey.14
For questions with levels of agreement or importance, the proportion of valid responses for each level was determined. Mean ± SD were ascertained, as appropriate. Ratings for the ideal and actual importance of the mentor for training in specific skills areas, from the perspective of the K recipient and the mentor, were compared using the Pearson’s chi-square test. A p ≤ 0.05 was considered significant. A nominal logistic model utilizing items from the Mentorship scale determined satisfaction with the mentor relationship. For purposes of analysis, the dependent variable responses were collapsed into two categories: satisfied (agreed or strongly agreed with “I am satisfied with my mentor relationship”), or dissatisfied (disagreed or strongly disagreed with that statement). JMP 9 (SAS Institute, Cary, NC, USA) was used for analysis of data.
Demographics for the recipients are shown in Table 1. Almost an equal percentage of women and men K recipients responded to the survey (52.0% and 48.0%, respectively). A total of 75.5% of the K recipients in our survey said their mentor was male. A total of 72.1% of mentor respondents were male. The mean age of the K recipients was 40.4 ± 5.2years. The mean age of the mentor was 56.0 ± 0.4 years. The average number of K recipients mentored was 1.9 ± 1.4. Many of these individuals also mentored postdoctoral fellows, and collectively the average number of K’s and non-K postdoctoral fellows mentored currently or in the past was 9.5 ± 9.1. K recipients had, on average, completed 3.1 ± 1.4 years of their K award program. The mentors reported that their most senior K recipient (the one they referenced to complete the survey) had completed 2.9 ± 0.1 years of their K award.
It should be noted that the ethnic and racial distribution included only 3.2% (29/901) of K recipients self-identifying as black and 5.1% as Hispanic. None of the mentors self-identified as black, but 94.9% identified as white with 3.2% reporting ethnicity as Hispanic.
The Mentorship Profile Questionnaire asked K respondents to report on outcomes that resulted from their interaction with the mentors. Research related outcomes were more common than general academic activities: presentation or poster (88.7%); publication (86.4%); grant writing or submission (85.4%); new research method or strategy (69.5%); job change or promotion (50.7%); service activity (32.6%); clinical experience (27.9%); and development of a program (24.8%).
Respondents were asked how important a mentor should be, in an ideal situation, for training the K recipient in specific research areas. They were then asked how important their mentor had actually been or how important they as the mentor had been in that area. Table 2 shows responses for both the ideal and the actual role of the mentor from the K recipient and mentor perspective. While 90.7% of recipients thought the mentor should be important or very important in training for research methods, almost 24.2% reported their mentor had been only somewhat or not at all important. Ideally, mentor input in working with the IRB or IACUC was thought to be important or very important by 73.7% of K recipients but 58.4% said that the mentor actually was only somewhat or not at all important. This area was rated with the lowest ideal importance for training by both K recipients and mentors. Except for networking, mentors tended to count the ideal importance of a mentor as very important more often than the recipients. For all areas, both K recipients and mentors rated the ideal importance of a mentor’s input as higher than the actual role for specific skill training (p < 0.0001 ideal compared to actual for both the K recipients and the mentors).
Interactions between the mentor and K recipient were assessed by determining the frequency of both in person meetings, as well as communication via email or phone. Table 3 shows the frequency of interactions as reported by K recipients and mentors. Both K recipients and mentors tended to note more frequent communication via phone or email rather than in person meetings and mentors more frequently noted daily interaction. The majority of K recipients and mentors reported interacting at least once a week. Fewer than 5% reported communicating less often than once a quarter.
The mentor qualities from the Mentorship Effectiveness Scale, as reported by the K recipients are shown in Table 4. Recipients primarily reported agreement with positive qualities of mentors. The lowest levels of agreement were seen with the mentor providing professional guidance and for challenging the K recipient to extend abilities.
Overall, 88.2% of recipients were satisfied with their mentors. Neither the K recipient’s nor the mentor’s gender were significantly correlated with satisfaction. However, the race of the K appears to be important to satisfaction with the mentor relationship. While the numbers were small, black respondents were more likely to be dissatisfied with the mentor relationship (6/29 or 20.7% of black respondents) than either their white (91/691 or 13.2%) or Asian (5/135 or 3.7%) counterparts. The K recipients were not asked the race/ethnicity of their mentor.
Interaction with the mentor contributed to satisfaction with the mentor relationship. In person meeting and communication by either email or phone was more frequent in those who were satisfied than those who were dissatisfied. (Meeting frequency of satisfied compared to not satisfied: N 879, Pearson’s chi-square 167.2, p < 0.0001; communication frequency satisfied compared to not satisfied: N 875, Pearson’s chi-square 169.1 p < 0.0001). Refer to Figure 1 and Figure 2, which show the frequency of communication and meeting by the K recipients who were satisfied or dissatisfied with their mentor relationship.
Nominal logistic modeling was used to examine the effect of the mentor qualities from the Mentorship Effectiveness Questionnaire on the satisfaction with the mentorship relationship. The model demonstrated significance (N 848, R2 0.648, p < 0.001), with the following variables: mentor accessibility, content expertise, support, guidance, recognition of contributions, and constructive input. Table 5 shows statistics on the individual components of the model.
While the response rate to our survey was somewhat low (approximately 31% of K’s and 32% of mentors), it is similar to other Web-based surveys. For example, a study examining response rates across 39 studies that used Web-based surveys found an average response rate of 34%. 15 To assess the representativeness of our respondents to the population of K’s and mentors, to the extent possible, we compared them to the national profile of K recipients and mentors. Because reporting of such demographics as age, gender, and race is voluntary on grant applications, K recipients who responded to this study cannot be compared directly to known demographics of all K recipients. The gender representation of the K recipients in our survey is similar to NIH data indicating that approximately 45% of research career award recipients in 2008 were women.16 The average age of our K recipient respondents (40.4 years) is quite consistent with NIH data on the average age of first time R01 principal investigators (42.6 years).17 Given that our K recipients had completed on average 3 years of their K program, their age at survey completion is also consistent with the recent report on K01, K08, and K23 recipients showing a median age at award of 37 years for all applicants to these award programs.18 The middle age of the mentors was not surprising considering these were most likely senior investigators. Although there is no existing data on K mentors to compare to our respondents, as a point of reference, in 2008 women accounted for approximately 28% of NIH research awards15 which is very similar to reported gender of the mentor by the recipients and the representation of female mentor respondents in our survey.
The National Institutes of Health Individual Mentored Career Development Awards Program report used several databases to report on the race/ethnicity of K applicants. It excluded K01 programs that were specifically targeted to diversity. For the remaining 9403 K01 (during 2000–2005), K08 (during 1990–2005), and K23 (during 2000–2005) applicants, 9.8% had unknown or unreported race/ethnicity. Of those that were reported, 2.6% of the K applicants self-identified as black and 3.4% as Hispanic. 18 Our survey did not exclude K01’s targeting diversity. Assuming that the respondents in our survey were similar to the entire pool of K awardees, there were several important findings. Perhaps the most concerning was the ethnic and racial makeup of the respondents in our sample, with Hispanics comprising 5.1% and blacks 3.2% of the total respondents. The ethnic and racial makeup of the mentor sample, with Hispanics represented by only 3.2% of the mentors and no black mentors, parallels the K recipients. Even presuming a response bias, these low percentages reflect a need to evaluate and promote programs that encourage minority participation of investigators and mentors in research.
A recent study by Ginther et al. evaluated R01 research project grants submitted between 2000 and 2006 by Ph.D. investigators at US institutions, including 83,188 applications and 40,069 unique investigators. Twenty-one percent of the applications were from Asian, black, Hispanic or Native American investigators. Of those only 1.4% were from black investigators. In addition to low numbers of black applicants, NIH R01 applications from black investigators were 13.2 percentage points less likely to be awarded (p < 0.001) than applications from white investigators. 19 Tabak and Collins further evaluated the funding success rates of minorities by looking at the success rate of underrepresented minorities who had previously received the National Institute on Minority Health and Health Disparities (NIMHD) Loan Repayment Program (LRP) or an NIH Research Training and Fellowship Award (T and F series), or a Career Development Award (K series). They found a similar trend of lower success rates and suggested that Ginther’s results might be generalizable beyond R01 funding. 20 They also noted that blacks and Hispanics were less likely to reapply after failing to receive funding on their first try. Establishing a core of minority mentors will be difficult unless the number of funded minority investigators increases.
Progress to encourage and increase minority and women’s participation in research will take time to build a sufficient minority and female senior investigator pool who can be mentors. While this diversity is important in the overall research enterprise, it may not be as important in the development of minority of female researchers since having a mentor appears more important than the mentor being from the same ethnic or cultural group. 3 A 1995 survey of full time medical school faculty reported that 86% of minority faculty did not think that it was important to have mentors of their same gender or minority group.4 In a 1998 survey of instructors and assistant professors at Harvard Medical School, accounting for ethnic and cultural issues and gender was not important to satisfaction with mentoring. 21 In contrast to the above two studies, this study asked K recipients directly about their satisfaction with their mentor relationship. Although overall 11.8% of K’s were dissatisfied with their mentor, it is concerning that 20.7% of black K’s were dissatisfied. How and whether race contributes to such dissatisfaction is not known. It must be kept in mind that there were only 29 black K respondents in our sample, which again points out a need to further evaluate minority participation in the K program. Although there were fewer women than men service as mentors, the gender of the K recipient and the mentor did not significantly correlate with satisfaction in our study.
Among this study’s K recipient respondents, 46.4% held Ph.D. and 51.1% M.D. degrees. A total of 49.6% of the mentors held Ph.D. and 48.3% M.D. degrees. The NIH Extramural Nexus recently reported that approximately 70% of NIH supported PIs have Ph.D.’s which has remained stable over the past 25 years. Most of the remaining PI’s were M.D.s or M.D./Ph.D.s with small percentages in other degrees. The NIH also reports that the percentage of M.D.’s receiving NIH awards has declined from approximately 20% to 17%. 22 Such a decline is concerning in light of increasing NIH emphasis on clinical translational research. In this study, more of the degrees held by the K recipients and mentors were M.D.s (approximately 50%) as compared to the NIH report. This difference is likely due to the clinical focus of these K awards. It is encouraging that M.D. investigators are serving as mentors for these clinically oriented K recipients. It is also interesting to note that 20% of K awardees had industry funding and 40% had funding from nonprofit sources other than NIH. Having experience with diverse funding sources is likely to serve them well in an era of declining (or flattening) of NIH budgets. Kram has classified mentoring activities into two kinds: instrumental and psychosocial. Instrumental support is influential in furthering the mentee’s professional goals, including reviewing and writing manuscripts and grant proposals, discussing research projects, and providing sponsorship, coaching, exposure, opportunities, and challenging assignments. Psychosocial mentoring includes role modeling, empathizing, and counseling as well as supporting the scholar through difficult courses and challenging obstacles. 23 Satisfaction with the mentor relationship in our study was consistent with satisfaction with both instrumental and psychosocial types of activities (see Tables 4 and and55).
Our study shows that mentors spend time in discussions, meetings, and email exchanges with their mentee. The frequency of contact either in a meeting or by phone/email was significantly correlated with mentee satisfaction in the relationship with their mentor (Table 3, Figures 1 and and2).2). This correlation has been shown in other reports of mentee–mentor relationships in general academic mentoring. 24
The assessment by a mentee of the mentoring experience is related to mentee outcomes. In a study of primary care fellows, those fellows who had “influential” mentors spent more time conducting research and published more papers. When asked about the attributes of their mentors, the respondents identified three domains: the relationship between themselves and their mentor (e.g., advice and guidance); the personal attributes of their mentor (e.g., kind and thoughtful); and the professional attributes of their mentor (e.g., role model). 10 Given that the K mentor relationship is intended for research, it is not surprising that the outcomes reported by the K’s in this study are higher in research related areas (presentations or posters, publications, and grant writing) then other areas of professional growth.
Having a mentor in the early years has been cited as a critical element in starting a productive professional career.25 In academic medicine, studies have shown that faculty members who identified a mentor felt more confident than their peers, were more likely to have a productive research career, and reported greater career satisfaction.4,6,11,26,27 Only one study about K recipients’ perspectives on their mentors was located. 28 This study used focus groups (n = 28) with K23 grantees to better understand the benefits and challenges they faced in the grant program. The discussion section noted that 26 of the 28 respondents indicated that their mentor “fostered involvement in research and scholarly activity.”27
The NIH Clinical and Translational Science Awards (CTSA; (RFA-RM-09–004) are designed to support model academic homes that provide resources and promote clinical and translational research that is aimed to improve health and prevent disease. 29 Mentoring of new investigators is a vital part of these programs. The grant requires institutions to provide not only a mentoring pool but also plans for training and evaluating the mentors, or replacing them, if necessary. In 2009, only 13 of the 46 CTSA institutions had formal mentoring training programs. 31 This same study of CTSA centers with KL2 scholars (a CTSA institutional mentored career development program to develop clinical research leaders), showed that only 17 of the 46 KL2 programs currently conducted formal evaluations of the mentoring relationship and another five were developing an evaluation instrument.30 Some of these institutions utilize the Berk et al.2 tool used in our study. Future comparison of data among CTSAs regarding mentoring qualities and satisfaction may help clarify the attributes of an effective research mentor relationship.
As previously noted, K awards do not provide salary or other support for the mentors yet they are expected to not only share knowledge but also their research resources and time. Burnham et al. have stated, “There is a pressing need for the research community to address the workload, institutional expectations, and reward system for research mentors.”31 Burnham and colleagues recommend several methods of supporting research mentors including monetary support, access to core laboratories and other CTSA-related services, assistance with grant and manuscript preparation, membership in mentor academies, mentor awards, training, promotion, and peer support. Identifying and implementing additional supports for research mentors is a particularly important issue if we also expect them to participate in mentor training programs. Essential to this purpose will be continuous evaluation of mentor supports and their impact on the outcomes we want to achieve from high quality mentoring relationships.
This study focused on a very specific type of new investigator, a member of an elite group with success in obtaining NIH funding providing both research and training support. The demographics of our study cannot be compared to all K recipients and their mentors, although there is little doubt that minorities are underrepresented in both and women are under-represented as mentors. It is also disconcerting to see the high rate of dissatisfaction with the mentor relationship found among the black K recipients responding to the survey. While current efforts appear to be directed at bringing more individuals from under-represented groups into the research pipeline, there should be additional consideration given to identifying, training, and supporting individuals who can serve in key mentoring roles for these individuals.
Both K recipients and their mentors reported that a mentor should be important to the provision of training in specific areas of research, yet neither thought that the mentor or they, themselves, had completely fulfilled this role. Determining why mentors are not as effective in training as both mentors and mentees thought they should be needs further evaluation. Specific mentor training that addresses these factors may benefit both the mentor and mentee.
Given that having a mentor is associated with increased productivity, developing strong mentor relationships is important. This study shows that there are mentor qualities associated with a high satisfaction with the mentor relationship. These qualities include more frequent interactions, accessibility, content expertise, offering support and guidance, recognition of their trainees’ contribution to work, and constructive input. Although mentors are required, K funding announcements do not give specifics regarding the mentor’s amount or content of contributions. As organizations, and especially CTSA’s, develop mentoring guidelines and training, more specific guidance should be given to these new investigators in choosing and building a successful mentoring relationship.
This study was funded by Grant 5R21RR025262-02 and UL1RR031990 from the National Center for Research Resources (NCRR) and NIH Roadmap for Medical Research, National Institutes of Health.