We retrieved 3625 total citations in PubMed, ERIC, PsycINFO, and the Cochrane Database of Systematic Reviews in June 2017. After screening for duplicate records, 3308 remained. An additional reference, which had yet to be indexed in PubMed, was identified outside of the search for a total of 3309 references. A total of 2577 records were excluded based on title and abstract screening, and 732 underwent full-text review (Fig. 1). Twenty studies met all of our inclusion criteria and were included in our review. These 20 studies described 19 separate mentorship programs. One article described two separate programs at two different institutions.12 Three articles reported results from the same program in pilot form,13 an expanded program,14 and with long-term outcome data demonstrating objective outcomes in all three phases of the program.15 The results of these articles are presented and listed in Table 1.
Twelve of the 19 programs were designed for junior faculty, as defined by individual study authors.12,13,14,15,16,17,18,19,20,21,22,23,24 While most of the included programs were exclusively for medical doctorates, one program for junior faculty was interdisciplinary and included female psychologists and social workers.17 Four programs included physicians across a range of levels25,26,27,28 and three of the programs included only those at the trainee level: two for medical students29, 30 and one for residents.31 Fifteen of the programs were specifically developed for female physicians,12,13,14,15, 17,18,19,20,21, 24,25,26,27,28,29,30,31 and ten programs used exclusively female faculty as the mentors.12,13,14,15, 18, 20, 24, 26,27,28,29,30,31 In five of the programs, the gender of the mentors was unclear16,17,18, 22, 23, 25 and the remaining programs included both male and female mentors.12, 19, 21 Ten of the programs described larger institutional professional development initiatives for women for which mentorship was a component of the initiative.12, 16, 17, 20, 25,26,27,28, 30, 31 Commonly described components of these larger programs included workshops for participants, development of electronic resources, classes for skills development, networking events, guest speakers, and journal and/or book club discussions.
The most commonly cited model of mentorship was the traditional dyad model (i.e., one mentor/one mentee), which was used by ten of the programs.12, 16, 19,20,21,22,23, 27, 29 Peer and facilitated peer mentoring was the next most commonly cited model, used in four of the programs.13,14,15, 17, 18, 24 Group mentorship was used in two of the programs,28, 30 and for the remaining programs, the mentorship model was not clearly delineated.25, 26 Further details on the definitions of peer, facilitated peer, and group mentorship have all been previously described in the literature.32
While many of the programs described broad objectives related to career development, faculty development, or mentorship and guidance, several of the programs had more specific objectives aimed at faculty recruitment, retention, promotion, and scholarship. Four of the 20 programs specifically aimed to increase the number of women within the specialty or department: two programs were intended to recruit female medical students into specific specialties (surgery and diagnostic radiology)27, 29 and two were designed for female faculty in specific departments (ENT and psychiatry).17, 26 Six of the programs set objectives related to the promotion of female faculty.12, 16, 20, 21, 25 Three programs were focused on improving scholarship13,14,15,16, 18, 22 with two additional ones aimed at improving specific skills important to a career in academic medicine, such as public speaking or teaching.27, 31
The majority of the programs’ primary outcome was participant satisfaction with the program, as measured by surveys. The programs were consistently highly rated by the participants. Among the 14/20 programs that reported on gender concordance between mentors and mentees, no differences emerged among programs for which all mentors were female and programs for which male mentors were included.
Eight articles provided objective outcomes for their programs, citing increased recruitment, retention, or promotion of female faculty or publication of scholarship.13,14,15,16,17, 25,26,27 One program reported an 85% retention rate for the female faculty who participated in the program which was higher than the national average.16 Another program reported improvement in the number of female faculty and female department chairs, as well as improvement in retention with a departure rate of women from the school of medicine of 34%, compared to a national average of 40%.25 Regarding the papers that described the same program in three phases, for both the pilot and expanded programs, the studies reported on submitted manuscripts as an outcome.13,14,15 The paper describing long-term data reported on faculty promotion based on retrospective CV review, finding that 6 of the 16 faculty received a promotion up to 5 years after enrollment. Two other programs reported improvement in recruitment of women into the department or the specialty.17, 27
Few articles described the cost of the programs. Only one article provided a cost breakdown which showed that their program, which involved significant faculty time for attending weekly half-day workshops and cost $670,000 over 4 years including salary support for faculty members’ time to participate.16 However, they reported that given the improvement in the retention rate of their female faculty, the program produced a cost savings of $330,000 over 4 years. One program reported that their mentorship program was provided with financial support from the deans’ office of approximately $30,000 per year.12 Another program reported a small budget of $2500 provided by the residency office.27 An additional program did not report specific cost in terms of faculty time but reported that each peer mentee was provided with 25 h of protected time for which to attend the program.13
Limitations or Barriers of Programs
Very few of the articles discussed barriers of the mentorship programs; however, several limitations are evident. In general, the sample size was small, ranging from 413 to 9320 participants and none of the programs had a control group, which creates a concern for selection bias as those who chose to participate in a mentorship program likely differ from faculty that choose not to participate. Furthermore, a lack of a control group also limits the ability to draw definitive conclusions about the impact of a specific mentorship program, as other institutional initiatives may have occurred simultaneously that could have influenced results of these manuscripts. Additionally, the short-term follow-up, subjective outcome measures, and unclear response rates in many of the papers limit the ability to evaluate both the immediate and, importantly, the sustained impacts of these programs. Finally, some of the authors commented that, occasionally, the matched mentoring pairs did not result in ideal pairings, thus limiting the effectiveness of the intervention for some participants.22, 30
Among those programs that cited specific barriers, one cited a lack of support staff as a barrier to the success of the program.12 Two cited time as a limitation for both the mentor19 and mentees,24 and it is likely that all mentorship programs face this barrier to a significant degree. It is also difficult to quantify the degree to which time invested in mentorship may result in future time saved in optimizing productivity and effectiveness at work.