Mentorship Programs in Residency: A Scoping Review

The benefits of mentorship during residency training are well established.^1-5  Research has demonstrated a positive correlation between mentorship and professional development, productivity, and academic success.^1-5  Residents who identify mentors are more likely to be hired in their desired specialties and pass qualifying examinations.⁴  Mentorship is also valuable for resident well-being, increasing job satisfaction and reducing rates of resident burnout.^6-8  Furthermore, mentorship can advance diversity, equity, and inclusion in medicine, as those underrepresented in medicine may experience unique benefits from mentorship.^9-11 

Despite the established benefits of mentorship, reported access to mentorship during residency training varies between individual trainees and institutions.¹  Including formal mentorship programs in residency programs is a strategy to increase access to mentorship. Formal mentorship programs are reported to exist in 50% to 82% of residency programs in Canada and the United States, most of which have been implemented within the last decade.^12-15  Despite widespread implementation in residency programs, strategies to guide the development, implementation, and evaluation of formal mentorship programs have not previously been reviewed and synthesized. As a result, existing and future mentorship programs may fall short in their potential to meet residents' needs. We undertook a review of the literature to synthesize existing evidence on formal mentorship programs within residency programs. Scoping review methodology was utilized to broadly examine the literature and assess heterogeneity of existing studies, outcomes of interest, and knowledge gaps.¹⁶ 

The objective of this study was to synthesize the current literature on formal mentorship programs within residency programs in the United States and Canada to identify key barriers and facilitators for successful programs. Specifically, this study aimed to describe the structure of formal mentorship programs, characterize resident experiences and outcomes, and describe tools and strategies used to evaluate outcomes from formal mentorship programs.

We adhered to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) checklist for scoping review guidelines.¹⁷  Our review protocol is provided as online supplementary data.

The study included residency or fellowship programs in the United States and/or Canada. We focused on these countries due to the common structure of residency training and postgraduate medical education. To be included, studies had to address the design, structure, implementation, evaluation, outcome, or resident/fellow experience with a formal mentorship program(s). Since one of the overarching objectives of our study was to inform improvements to mentorship programs in residency, we included only studies which aimed to benefit residents (as opposed to institutions) through their mentorship programs.

Any study type was eligible for inclusion, including surveys, observational studies, qualitative studies, abstracts, and commentaries. Full inclusion and exclusion criteria are shown in Table 1.

We searched Ovid MEDLINE (between 1940 and 2019) and Embase (between 1980 and 2019). We also performed hand searching of the reference lists of all included studies. The most recent search was executed on January 2, 2020.

Our search strategy, included in online supplementary data, was developed in collaboration with a medical librarian and adapted from a previous systematic review by Pethrick et al.¹⁸ 

References were imported into Covidence and de-duplicated. Each abstract (or full article for those without an abstract) was independently screened by 2 authors for eligibility, and all disagreements were resolved by review from a third author.

A data extraction tool was developed by 2 members of the study team (M.B.J., K.L.) based on preliminary review of included articles. The extraction tool was piloted by all study team members and revised based on their feedback and its usability. Data from each included article were extracted in parallel and reconciled by 2 study team members. All differences were resolved by consultation of the full-text article, and disagreements were resolved by consultation with a third author.

We collected data on study type, setting, population, participant characteristics, program characteristics, and evaluation. A template of our data extraction tool with complete details of all collected variables can be found in the online supplementary data.

We performed critical appraisal for all peer-reviewed studies using the following tools according to the study methodology: The Critical Appraisal Skills Programme Qualitative Studies Checklist was used for interview and qualitative data,¹⁹  and the appropriate risk of bias instruments provided by McMaster University's CLARITY Group were used to assess surveys, cohort studies, case control studies, and randomized controlled trials.^20-23 

Furthermore, our data extraction tool included a field for the authors to enter free text regarding study strengths and limitations, which were discussed and collated at the time of reconciliation. Critical appraisal was performed to understand the strengths and limitations of the current literature but did not inform inclusion or exclusion of studies.

We performed descriptive numerical analysis for collected variables, including study, program, mentor, and evaluation characteristics. Additional analyses were based on common themes of interest that emerged during the full-text review and data extraction.

Results from our search strategy are shown in the Figure. Fifty-five studies fulfilled inclusion criteria and underwent data extraction by 2 authors (Figure). We did not identify any full studies via hand search of the reference list of studies identified by our initial search strategy.

Characteristics of included studies are outlined in Table 2. The earliest study was published in 1995. Four studies (7%) focused on specific mentee populations^24-27 ; 2 studies (4%) reported on mentorship programs and included women trainees only^24,25 ; one study (2%) described a program that was reserved for fellows enrolled in a clinician educator program²⁶ ; and one study (2%) examined a mentorship program for international medical graduates.²⁷ 

Eleven (41%) of the studies^24,28-37  that assigned mentors to mentees attempted to match residents with mentors on the basis of mentee characteristics, rather than random assignment. Strategies for matching mentors with mentees included the mentee's personal interests,^{24,28-31,34,35}  gender identity,^29,32,33,35  and career goals.^{24,29,30,33,34} 

Included studies reported conflicting findings for assigned versus chosen mentors. One study in a combined internal medicine and pediatrics program compared random matching of residents with mentors to matching based on compatibility scores (assessing demographics, personal interests, values, and professional goals) and found no significant differences in results of a program evaluation survey a year after implementation.³⁷  A study of 27 radiology residents found that residents who self-selected their mentors (by requesting a specific faculty member) had higher frequency of meetings and higher satisfaction with the mentorship program compared to residents who were assigned a mentor, even when mentors were assigned to residents based on their preferred attributes.³⁸  Another study of 204 psychiatry residents across 12 residency programs reported that residents with self-selected mentors were significantly more likely to agree that their mentor had a positive impact on research and scholarly activity compared to residents with assigned mentors.³⁹  However, in a survey of 179 residents from 17 general surgery programs, Delisle et al found there was no difference in scores on the Mentorship Effectiveness Scale based on whether mentors were chosen or assigned.⁴⁰ 

Mentor-to-mentee ratio was described in 37 studies (67% of total included), although no study compared the influence of these ratios on any outcome. Half of these studies examined a one-on-one, mentor-to-mentee ratio (49%, 18 of 37).^{26-28,31,32,34,36,38,41-50}  In 6 of these studies (16%), there was more than one mentor available for each resident.^33,36,51-54  In 7 of these studies (19%), a mentor had multiple mentees in a mentorship group.^30,35,55-59  The remaining 6 studies (16%) reported variable mentor to mentee ratios.^{3,29,37,60-62} 

Timing of meetings was reported in 15 studies (27% of total included). Eight of these studies (53%) reported that mentorship program meetings occurred during work hours,^* while 4 (27%) reported mentorship program meetings outside of work hours,^25,30,41,64  and 3 (20%) reported a mixture of meeting times.^24,32,60  Of the studies where meetings took place during work hours, only 4 (50%) reported scheduling protected nonclinical time for residents to attend these meetings.^44,46,47,50  Frequency of meetings most commonly occurred every 3 to 6 months^†; however, other studies reported meeting weekly,⁵³  every 1 to 2 months,^{26,46,55,58,59,64}  every 6 months,⁶⁰  or at variable intervals.⁶² 

Meeting location was reported in 5 studies (9% of total included). Three of these (60%) reported meetings took place at the hospital or teaching site,^46,50,55  one (20%) reported a mixture of meetings at formal (ie, mentor's office) or informal (ie, restaurant) settings,⁶⁰  and one (20%) reported meeting virtually via Skype.⁴¹  Of note, one (20%) study identified meeting in informal settings as a facilitator for mentorship.⁶⁰ 

Forty-two studies (76% of total included) reported on mentor characteristics. Of these, 34 mentorship programs (81%) utilized staff physician mentors,^‡ 3 (7%) used resident peer mentors,^35,44,55  and 6 (14%) used a mixture of peer and staff physician mentors.^{30,50,56,58,59,64} 

Recruitment of mentors was reported in 21 studies. Mentors were recruited on a volunteer basis^§ and/or invited to become mentors based on certain characteristics or credentials.^{31,33,36,58-61,66,69}  Thirteen of the 39 single-program studies (33%) reported on providing mentors with instructional resources on mentorship^‖ and/or formal mentorship training.^{55,58-60,71,73}  One survey of 80 neurological surgical programs noted that educational resources for mentors were provided in only 34% of programs.⁷⁴ 

Only 8 studies (15%) mentioned the issue of unsuccessful mentoring relationships and stated that there would be an option for residents to switch mentors if this occurred.^{28,29,36,45,65,68-70} 

Program goals were reported by 29 studies (53%) and included professional development,^# career planning,^{28,29,33,63,69,72}  scholarly activity,^{26,29,33,61,63,76}  improving wellness,^{29,33,36,46,59,63}  education,^{28,48,49,57,58}  providing psychosocial support,^{28,35,50,53,69}  providing networking opportunities,^24,33,63  reducing burnout,^36,46,55  aiding with transition to residency⁵⁰  or transition to practice,⁶⁸  and improving resilience.⁵⁵ 

There were 39 (71%) single-program studies that performed a formal evaluation of their respective mentorship programs.^** Further details of program evaluations are shown in Table 3.

Main findings from the 55 included studies are synthesized in Table 4. Most commonly, investigators used a survey at one time point to evaluate self-reported resident satisfaction with the program.^†† Due to heterogeneous and generally low-quality survey methodology based on critical appraisal, we were unable to synthesize these results. Only 4 of the single-program studies (10%) reported using the evaluation results to inform program changes.^29,56,66,69  Three (8%) used pre-existing instruments to study their programs; these included the Mentorship Profile Questionnaire,⁶¹  Mentorship Effectiveness Scale,^40,61,78  Perceived Stress Scale (PSS),^36,79  Maslach Burnout Inventory (MBI),^36,80  WHO Quality of Life-BREF (WHOQOL-BREF),^36,81  and Munich Evaluation of Mentoring Questionnaire.⁸²  Given the limited data, it was not possible to determine the efficacy of mentorship program designs between different studies.

Three single-program studies (8%) matched the method of evaluation to the program objectives. Chandler and Borum described a program which aimed to enhance medical knowledge, and which assessed scores on the in-training examination before and after program implementation.⁴⁸  They reported an increase in mean scores on the examination the year after implementation and an increase in the training program's percentile rank compared with national average scores.⁴⁸  Zhang and colleagues³⁶  aimed to reduce burnout and improve well-being and performed evaluation with tools including the PSS,⁷⁹  MBI,⁸⁰  and WHOQOL-BREF.⁸¹  At 12 months following implementation of their mentorship program, participants reported statistically significant reduction in PSS scores; improved scores on the emotional, levels of depersonalization, and personal achievement domains of the MBI; and improvement in overall WHOQOL-BREF scores.³⁶  Saint Martin et al⁴⁶  aimed to address burnout and improve wellness, and evaluated wellness knowledge and utilization of wellness techniques before and after implementation of their program. Twelve months after program implementation, they reported improved scores on a wellness survey and improved self-reported rates of personal wellness knowledge and utilization of wellness techniques.⁴⁶ 

The reporting of important methods details or results also limited our ability to compare studies. Across studies, the median number of participants was 40 (range 2 to 585). Participation rates were not reported in 14 of the 34 survey studies (41%).^‡‡ Of those studies that reported participation rates, the median response rate was 64% (range 14% to 100%).

Twelve studies explored barriers and facilitators to successful mentorship programs using surveys, focus groups, or interview data.^§§ Compatibility between mentors and mentees was a common theme; flexible and organic pairing⁶⁰  as well as common personal interests, personalities, and goals were identified as facilitators of successful mentoring relationships.⁷⁰  Accessibility of the program was another theme; geographic proximity of meetings and approachable and accessible mentors were identified as facilitators,^60,70  whereas scheduling and time constraints (for both mentors and mentees) were identified as barriers.^‖‖ Mentee characteristics including engagement, buy-in, and self-motivation were facilitators of successful mentorship programs.^58,70,74  A full list of barriers and facilitators from these studies is summarized in Table 5.

Three studies used inductive coding techniques to perform qualitative analysis of focus group or interview data obtained from mentees^60,70,76  and mentors.^60,70 

This scoping review examined studies that described formal postgraduate medicine mentorship programs and, when reported, their associated evaluations and outcomes. We found that mentorship programs varied with respect to their stated objectives, design, structure, and evaluation. We synthesized the described barriers to and facilitators for successful mentorship programs to identify common themes (Table 5).

Only 3 studies utilized high-quality evaluation tools tailored to their program objectives.^36,46,48  To best enhance knowledge on this subject and facilitate improvements in structure of formal mentorship programs, future studies in this field should tailor program evaluation strategies to program objectives.

Few studies reported on an option for switching mentors due to lack of compatibility. ^{28,29,36,45,65,68-70}  Previous literature has highlighted the importance of mentor-mentee compatibility.⁸⁶  However, due to the power differential, mentees may not feel comfortable asking for a new mentor or dissolving the relationship. As such, we suggest programs have a low barrier mechanism for mentees to switch mentors if the relationship with their initial mentor is unsuccessful, for example, a scheduled review with the mentorship program director.

Given that resident scheduling and time constraints were the most cited barrier to mentorship,^## facilitating protected time for mentorship meetings may increase accessibility. Other authors have emphasized mentor availability as a facilitator to mentorship.^87,88  Programs can evaluate accessibility by collecting data on number and/or frequency of meetings attended, and whether meetings were missed due to scheduling conflicts.

Unclear goals and expectations of mentorship programs were identified as barriers to mentorship,^60,70,76  and misalignment of goals and expectations between the mentor and mentee was reported to lead to negative mentorship outcomes for some residents.⁷⁰  Conversely, mentorship training and/or resources were identified as facilitators for successful mentoring relationships.⁷⁰  Previous literature suggests that clear communication of goals and expectations as well as formal instructional resources or training for mentors are facilitators of successful mentorship.⁸⁶ 

None of the studies specifically matched residents from historically excluded groups with mentors from similar demographics. Previous publications have reported that racial/ethnic concordance is of lower importance to mentees than other mentor characteristics⁸⁹  and does not adversely impact mentee satisfaction or success of a mentorship program.⁹⁰  We emphasize that, while residents from historically excluded groups likely benefit from mentorship,^9-11  it is unclear whether current programs narrow disparities by fulfilling unmet needs or widen disparities by continuing to underserve marginalized groups. Though this is beyond the scope of our review, we emphasize that principles of diversity, equity, and inclusion should be considered in design, implementation, and evaluation of formal mentorship programs.

The main limitation of our study is the quality and heterogeneity of evaluations performed by included studies, which limited our ability to synthesize outcome data. Another limitation is that the search period ended in 2019 so does not include more recently published articles. Our search covered 79 years (1940-2019). We included only English language studies performed in the United States and Canada, which may limit the generalizability of our findings.

To facilitate effective mentorship and overcome key barriers, the literature suggests programs (1) protect time for mentorship meetings; (2) set clear program goals and expectations; (3) provide mentors with resources and/or training; and (4) facilitate compatibility between mentors and mentees. Few mentorship program evaluations were tailored to stated program objectives, which represents a gap in this literature.