|Home | About | Journals | Submit | Contact Us | Français|
Mentorship is perceived to play a significant role in the career development and productivity of academic clinicians, but little is known about the characteristics of mentorship. This knowledge would be useful for those developing mentorship programs.
To complete a systematic review of the qualitative literature to explore and summarize the development, perceptions and experiences of the mentoring relationship in academic medicine.
Medline, PsycINFO, ERIC, Scopus and Current Contents databases from the earliest available date to December 2008.
We included studies that used qualitative research methodology to explore the meaning and characteristics of mentoring in academic medicine. Two investigators independently assessed articles for relevance and study quality, and extracted data using standardized forms. No restrictions were placed on the language of articles.
A total of 8,487 citations were identified, 114 full text articles were assessed, and 9 articles were selected for review. All studies were conducted in North America, and most focused on the initiation and cultivation phases of the mentoring relationship. Mentoring was described as a complex relationship based on mutual interests, both professional and personal. Mentees should take an active role in the formation and development of mentoring relationships. Good mentors should be sincere in their dealings with mentees, be able to listen actively and understand mentees' needs, and have a well-established position within the academic community. Some of the mentoring functions aim at the mentees’ academic growth and others at personal growth. Barriers to mentoring and dysfunctional mentoring can be related to personal factors, relational difficulties and structural/institutional barriers.
Successful mentoring requires commitment and interpersonal skills of the mentor and mentee, but also a facilitating environment at academic medicine's institutions.
The online version of this article (doi:10.1007/s11606-009-1165-8) contains supplementary material, which is available to authorized users.
Mentoring relationships have become an object of intense study, beginning with the seminal work by Kathy E. Kram in the 1980s,1 which initiated a surge of research in diverse settings such as business,2,3 education4,5 and nursing.6 In academic medicine, mentoring was recognized as a crucial developmental relationship,7 and our recent systematic review showed that mentoring has an important influence on personal development, career guidance, career choice and research productivity.8 The review we initially performed8 included only quantitative studies and focused exclusively on outcomes of mentorship. It did not address the meaning of mentoring, its formation and characteristics of its actors, which are more appropriately explored by the use of qualitative research methodologies.
Terms such as “mentoring,” “supervision” and “role modeling” can all be considered as describing developmental interactions, but they are often used interchangeably or without clear demarcation, which makes their operationalization more difficult. This lack of clarity has consequences in practice-oriented disciplines, where the development of programs is based on assumptions about the meaning and effectiveness of mentoring.5,9
Expert panels have made attempts to clarify the concept of mentoring in academic medicine,10,11 and authors have compiled narrative literature reviews12–14 or offered personal views15 on the desirable characteristics of mentoring relationships. Such approaches would be enriched by exploring the personal experience of mentoring as obtained through qualitative research methodologies. Acknowledging that the way people give meaning to concepts is context-dependent,16 we conducted a systematic review of qualitative research to explore what is known about the characteristics and dynamics of mentoring relationships in the context of academic medicine.
We sought studies that used qualitative research methods to explore the meaning and characteristics of mentoring relationships in academic medicine. We defined “qualitative research” as any study that placed people's own voices at the center of the interest and “academic medicine” as exploring a population of medical students or physicians at a medical school, university hospital or academic general practice. Excluded were quantitative studies as well as studies dealing with peer-mentoring, online- and tele-mentoring, and studies exploring short-term, task-oriented relationships such as clinical, research or educational supervision. The focus of this review was on a traditional, dyadic model of mentoring (one-to-one, senior-to-junior, face-to-face),17 which was the most frequently used model in formal mentoring programs described in the medical literature.9 We critically appraised all the included studies to identify their strengths and limitations, but we did not exclude any article based on insufficient methodological quality to avoid eliminating data germane to the purpose of the review.18–20
We searched the following databases: PsycINFO (1967-December Week 2, 2008); ERIC (1965-December Week 2, 2008); Ovid MEDLINE(R) (1950-November Week 4, 2008), Scopus, an Elsevier abstract and citation database (1996-November 28, 2008) and Ovid Current Contents/All Editions (1993–2008 Week 49). We used a combination of key words describing population, setting, target phenomenon and methodology21 (Appendix 1, online). There were no language restrictions.
Articles selected for inclusion were used as a starting point for the berry-picking search,22 which consisted of footnote chasing, citation searching, author searching and searching of “related articles” in MEDLINE. Fifteen authors of included articles were contacted to identify other potentially relevant articles. Reviewers' own work and knowledge of the literature, as well as the results of search for an earlier systematic review on mentoring8 were also used to find possible articles for inclusion.
Titles and abstracts of all retrieved articles were independently screened by two reviewers to determine if they met inclusion criteria. Full texts of the articles were reviewed by two reviewers (DS and SES), and in cases of disagreement, the third reviewer was consulted and a decision was made by consensus. Quality assessment and data abstraction were also completed independently by two investigators (DS and SES).
In the analysis of primary study findings we used the method of qualitative meta-summary,21 which refers to the non-interpretive aggregation of qualitative research findings. First, we extracted the findings by separating researchers' interpretations of primary data, which we considered as findings from other parts of the text, such as descriptions of data analysis procedures, quotations used to illustrate and support researchers’ interpretations, or researchers' discussions of the findings. Second, we edited the findings to allow better comprehension and grouped them into thematic categories. Third, we abstracted the findings to refine them and eliminate redundancies. Simultaneously with this process, we looked for possible ways to subclassify the categories, as well as for links and patterns that would allow more comprehensive understanding of the target phenomenon. We did not calculate manifest frequencies and intensity effect sizes,21 as there were too few articles to obtain meaningful information. Finally, for the categories that contained enough primary study findings, we developed taxonomies to show their conceptual range.
The analysis of findings was initially done by one of us (DS), and a draft of the meta-summary was shared with the other two authors, who suggested other possible approaches and interpretations. The discussion among the authors continued in the process of “negotiated consensual validation”21 until an agreement was reached.
The findings of primary studies are presented as statements with their respective reference numbers. All other statements in the Results section, related to grouping, classification or summarizing of primary studies’ findings, are made by the authors of the review.
A total of 8,487 citations were retrieved from the bibliographic database search, 114 full-text articles were reviewed, and 8 met inclusion criteria.23–30 The “berry-picking search” yielded 3,431 potentially relevant articles. Four full text articles were retrieved, and one of these met the inclusion criteria.31 Thus, the the total number of included articles was nine (Appendix 2, online). No additional relevant articles were identified by experts in the field.
The median (range) number of participants in the included studies was 18 (2–71), and all of the studies were conducted in North America. The nature of mentoring was not described in three studies,24,30,31 four studies explored formal mentoring relationships,23,26,28,29, one study explored informal mentoring relationships27 and one study provided participants with a definition of mentorship but did not describe if this was a formal process.25 Most of the studies included participants of both genders. More than half of the studies included both mentors and mentees, while the rest included only mentees. The majority of samples were self-selected or purposive/theoretical, and the data were mostly collected by interviews or focus groups. The analysis was mostly done by either thematic analysis or using a grounded theory approach (Appendix 3, online).
Quality assessment of the articles (Appendix 4, online) revealed that most studies had clearly stated objectives, but the description of the sample and sampling procedures sometimes lacked detail.23,24,27,28,31 Study findings were stated with varying levels of detail, and in one report it was difficult to discern the findings of the qualitative analysis.23
We identified five major themes in relation to mentoring relationships: (1) desired characteristics and actions of the mentor and mentee, (2) initiation of mentoring relationships, (3) structure of mentoring relationships, (4) characteristics of mentoring relationships, and (5) barriers and possible solutions to mentoring.
Four studies24,29–31 reported findings about the role of mentees. Mentees should take the initiative for cultivating the relationship with their mentors24 (taking the driver's seat31). It is important that mentees have commitment to the success of the mentoring relationship and passion to succeed in their career.31 Mentees must be proactive, willing to learn, and be selective in accepting advice from the mentor.30
It was perceived that for successful relationships, mentees should prepare for the meetings with their mentors, provide an outline of their activities for discussion, complete tasks that were agreed upon and respond honestly to feedback.29 Mentees should also perform self-reflection and reveal flaws so that their mentors can interpret and critique behavior.29 Courage is needed on the part of mentees to face their weaknesses and to make effective changes.29
Six studies24–26,29–31 reported the desired characteristics of the mentor. We classified these characteristics as pertaining to the mentor's personality, interpersonal abilities and professional status (Table 1). Good mentors should be honest29,31 and sincere24 in their dealings with mentees, be able to listen actively30 and understand mentees' needs,26,31 and have a well-established position within the academic community.25
Six studies24,25,27–29,31 explored the role of good mentors and how they interacted with their mentees. Some of the actions were aimed at academic growth of mentees, while others were targeted toward personal growth. We conceptualized this range of actions as a continuum, with an increasing level of privacy and intimacy from the institutional to the personal side (Table 2). On the institutional side, mentoring actions aimed to enhance mentees' visibility25 and connections within the academic environment25,29,31 and to protect them from adverse influences and harsh interactions.25 On the personal side, mentors helped to create a safe environment for expression of thoughts and feelings.29 They provided moral support24 and offered guidance in the processes of self-reflection,29 vision-building24,25 and goal-setting.29
Only two studies24,25 reported findings about the initiation of the mentoring relationship. These studies reported that emerging mentees had the responsibility to find a mentor,24,25 but their institutions could provide them with early guidance and education.24
Sometimes it was necessary to look for mentors in many places (inside and outside the department and institution, among peers and more senior faculty members).25 It was critical to locate a mentor early in one's academic career, but persistence and patience might also be necessary in finding a mentor.25
Mentoring can develop informally, as a relationship that evolves naturally over time toward mentoring commitments, or formally, as an assigned mentoring relationship.25 The former may be also termed as self identification of mentors by mentees. Participants in four studies25,26,30,31 raised concerns about the formal assignment of mentoring pairs. This assignment could ignore the interpersonal aspect (“chemistry”) of the relationship and thus prove less effective.25 Assigned mentorship could even have a negative impact, e.g., by making mentees feel “forced” into the relationship.31 Successful mentoring can develop through formal assignment, but it depends on the individuals involved.26 These same studies also suggested that self identification of mentors was perceived in many cases to be beneficial, allowing a more comfortable and effective relationship to develop.25,26,30,31
Structure relates to the gender/race/ethnic composition and the number of actors in the mentoring relationship. Experiences that members of different gender, race or ethnicity groups bring to their encounter can be seen as opportunities that allow for greater mutual growth.25 However, the findings about the need of gender/race/ethnic congruence, reported in five studies,25–27,30,31 were inconclusive. It seems that basing matches on such considerations was not essential25 and that the sensitivity of the mentor was more important than matching on any of these factors.26
One study found that female and minority residents and faculty, particularly those with, or intending to have, children, felt that gender is vital to the mentoring relationship.27 Concerns were raised in another study that a male mentor may have difficulty in giving criticism to women.26 Furthermore, a study found that male mentors were not always perceived to be able to provide guidance on the needs of a female mentee, especially in relation to the work and life balance, e.g., timing of maternity leave31 or managing a career while raising a family.27 The experience of women in academic medicine was perceived to be different from the experience of men, and therefore one study suggested that female mentees may benefit from the advice of another woman.26 Two studies identified the concern that the mentoring relationship can be tainted by sexual harassment, and it is important to set clear boundaries to prevent it.25,26
The findings regarding mentoring of minorities were even more tentative than those regarding mentoring women. Sensitivity in mentoring minorities is important, but the pairing by race or ethnicity, although preferred, is rarely possible.26
Systems of multiple mentors may be an appropriate answer to the challenges of gender, racial, ethnic or other differences that can make finding common ground in the mentoring relationship difficult.26 Participants in three studies25,26,31 acknowledged that many people might fill the mentoring role. Two models of dual mentorship were suggested: (1) a female mentee with a male mentor for her interest area and a female mentor for lifestyle issues (as there are usually not enough senior women to mentor junior women faculty in their interest area)26 and (2) academic mentor (needs to be local) for guidance on promotion, career milestones, local politics, work and life balance, and a scientific mentor (who can be at a distant site) for guidance on research.31 A disadvantage of multiple mentors is that they may have different opinions about the best course of the mentee's action and thus make the mentee's decisions more difficult.25 Good communication across mentorship teams is essential to avoid confusion.31
Personal connection Mentoring relationships were described as being based on professional and personal interests.30 They could be as complex and personal as the relationships with friends and family.25 They implied an exchange of information that allowed the mentor and the mentee to appreciate the other as a whole person.24 Mentoring relationships were potentially enhanced by similar interests and ideals (“chemistry”25 or “resonance”30) and challenged by differences.25 Both the mentor and the mentee should be able to recognize the changes in their relationship over time, with possible evolution into a peer relationship.26
Underlying values It was perceived that mentoring should be based on honesty,29 trust,29 mutual respect,29,31 open communication31 and confidentiality,31 all of which build an environment that is safe for self-exploration.29 Furthermore, there should be a willingness to take risks and commitment to resolve conflict.29 Both parties should acknowledge that the relationship centers on the needs of the mentee and make sure that the mentee clearly benefits from the relationship.29 Expectations from both should be made clear, particularly regarding what intellectual property belongs to the mentee.30
All reviewed studies reported barriers to mentoring. These could be classified as related to personal characteristics of the mentor or the mentee, their relationship, or institutional constraints and limitations (Table 3.) The same taxonomy was applied to possible solutions, or ways to improve mentoring, which were suggested in five studies23,24,28,30,31 (Table 4).
This systematic review of qualitative research identified a set of mentoring functions that provided psychosocial and career-related support.1 Mentors helped mentees flourish in the challenging environment of academic medicine by offering them emotional and moral support, working to build their personal and professional abilities, and providing them with backing and protection in their academic institutions. Personal inadequacies and relational problems were identified as the main barriers to mentoring, but structural constraints such as lack of time or incentive sometimes hindered the development of functional mentoring relationships.
Research in organizational settings has shown that mentoring can be distinguished from other developmental relationships such as leadership32 or coaching33 by the broadness of functions it offers. The range of desired actions and characteristics identified in this review indicates that mentoring in academic medicine is perceived as a uniquely encompassing relationship. Considering some of the mentoring actions described (e.g., helping the mentee to clarify feelings, motivating and fostering self-reflection), we conclude that a high-quality mentoring relationship, characterized by a high level of personal involvement and commitment, is meant to affect not only the professional, but also the personal lives of the mentor and mentee. The relational and reciprocal outcomes such as personal growth, interdependence and connectedness34 invite further exploration, especially as these types of outcomes of mentoring in academic medicine have been underinvestigated.7
In her classification of relational problems in mentoring, Eby35 envisions a continuum from minor problems, which presumably occur more frequently, to taxing ones, which are less frequent and more serious. We developed a different taxonomy consisting of personal, relational and structural dimensions, which enabled us to join barriers to mentoring and different types of dysfunctional mentoring in a single category. This seemed appropriate as there were too few examples and descriptions of relational problems in the studies we reviewed to produce a distinct classification of dysfunctional mentoring. Moreover, the three-dimension taxonomy was applicable to the suggested strategies of improving mentoring and allowed juxtaposing barriers to mentoring with ways to overcome them.
There are some limitations to this review. First, we did not search the grey literature, such as working papers or conference proceedings that were not formally published or subjected to peer review. However, we did a comprehensive search of available databases and contacted experts in the field. Despite our efforts, it is possible (as with any systematic review) that we did not retrieve all relevant studies. Second, all included studies have been conducted in North America, but this reflects the state of the existing research. Apart from the disproportionate geographical representation, the largest gap in the existing body of research relates to the limited depth in which the phenomenon of mentoring in academic medicine has been explored. In most of the included studies, authors performed a thematic analysis of mentoring experiences as reported by the participants, without providing a “thick” description of events and circumstances pertinent to mentoring, e.g., by doing an ethnographic study,36 or to develop a comprehensive theory, e.g., by doing a grounded theory study.37 This reduced our ability to gain a multifaceted and in-depth understanding of the mentoring relationship in academic medicine. Furthermore, it limited the level of conceptual innovation we could achieve in our systematic review.
Samples in the studies included in this systematic review had a balanced representation of male and female participants, majority and minority groups, as well as mentor and mentee perspective, thus providing a broad range of experiences. And our systematic review highlights areas for further qualitative research. Studies dealt mostly with the initiation and cultivation phases38 of mentoring in academic medicine, whereas the separation and redefinition phases38 were not well explored. These later stages in the life cycle of a mentoring relationship are important not only for the members of individual mentor-mentee dyad, but may also have broader consequences for the mentoring culture in academic medicine institutions. A more in-depth exploration of mentoring functions may be needed to inform the actions and interventions aimed to enhance mentoring. For example, the studies reported that mentors facilitated the mentee's visibility and exposure in the academic community. However, there were no research findings on how this is actually done, which actions and behaviors mentors perform to achieve this goal, and what the facilitators and barriers are in this process.
Mentoring in academic medicine is a complex phenomenon that affects the personal and professional lives of both mentor and mentee and implies a degree of intimacy that makes it resemble relationships with friends and family. Mentoring is inextricably situated in a social context and shaped by the institutional culture and climate. Successful mentoring therefore requires both commitment and interpersonal skills of mentor and mentee, and also a facilitating environment of the academic medicine's institutions. Academic institutions interested in developing mentorship strategies should respect the nature of mentoring relationships, which encompass professional, career-related and private aspects. Future research should focus more on the separation and redefinition phases of mentoring relationship and use a variety of qualitative methodologies and approaches to expand and deepen the body of knowledge about this important, yet elusive phenomenon.
We thank Ms Laure Perrier for her help in designing search strategies and retrieval of articles. Dr Sambunjak and Dr Marušić are financially supported in part by a grant from the Croatian Ministry of Science, Education and Sports (no. 216-1080314-0245) to Matko Marušić. Dr. Straus was supported by a Canada Research Chair.
Conflict of interest The funders had no influence on the design and conduct of the study; collection, management, analysis and interpretation of the data; or preparation, review or approval of the manuscript. All of the authors are teachers at medical schools and have no competing interests.