Abstract
Objective To collect information about the types, frequency, importance, and quality of mentorship received among academic family medicine faculty, and to identify variables associated with receiving high-quality mentorship.
Design Web-based survey of all faculty members of an academic department of family medicine.
Setting The Department of Family and Community Medicine of the University of Toronto in Ontario.
Participants All 1029 faculty members were invited to complete the survey.
Main outcome measures Receiving mentorship rated as very good or excellent in 1 or more of 6 content areas relevant to respondents’ professional lives, and information about demographic and practice characteristics, faculty ratings of their local departments and main practice settings, teaching activities, professional development, leadership, job satisfaction, and health. Bivariate and multivariate analyses identified variables associated with receiving high-quality mentorship.
Results The response rate was 66.8%. Almost all (95.0%) respondents had received mentorship in several areas, with informal mentorship being the most prevalent mode. Approximately 60% of respondents rated at least 1 area of mentoring as very good or excellent. Multivariate logistic regression identified 5 factors associated with an increased likelihood of rating mentorship quality as very good or excellent: positive perceptions of their local department (odds ratio [OR] = 4.02, 95% CI 2.47 to 6.54, P < .001); positive ratings of practice infrastructure (OR = 1.86, 95% CI 1.23 to 2.80, P = .003); increased frequency of receiving mentorship (OR = 2.78, 95% CI 1.59 to 4.89, P < .001); fewer years in practice (OR = 1.93, 95% CI 1.19 to 3.12, P = .007); and practising in a family practice teaching unit (OR = 1.51, 95% CI 1.01 to 2.27, P = .040).
Conclusion With increasing emphasis on distributed education and community-based teachers, family medicine faculties will need to develop strategies to support effective mentorship across a range of settings and career stages.
Mentorship is a vital component of faculty development in academic medicine and integral to growth in our roles as clinicians, teachers, leaders, and scholars.1 Much of the academic medical mentorship literature describes the nature of effective mentoring programs and roles, and the characteristics of successful mentoring relationships, both formal and informal.2–6 Straus and Sackett’s review of the research on mentorship in academic medicine revealed moderate-quality evidence that academic clinicians who are mentored report greater career satisfaction, are more productive, are promoted more quickly, and are more likely to stay at their institutions.7
However, few studies describe mentorship activities in academic family medicine departments.8,9 As academic family medicine has evolved, with changes that have affected the distribution, support, role expectations, and career satisfaction of faculty, understanding the perceptions and experiences of mentorship in this group is important.10–12 The purpose of this study was to identify the types, frequency, perceived importance, and quality of mentorship received within a large academic family medicine department, and to identify variables associated with receiving high-quality mentorship.
METHODS
Setting
In 2011, a work-life and leadership survey was sent to all faculty members of the Department of Family and Community Medicine (DFCM) at the University of Toronto in Ontario, the largest department of family medicine in North America. At the time of the survey there were 1029 faculty members, 248 postgraduate trainees, 226 clinical clerks, 26 faculty members with protected research time funded by the DFCM, and more than 20 family practice teaching units, with clinical teaching and research also taking place in dozens of community-based practices and clinics.
Questionnaire development and dissemination
The questionnaire was developed after reviewing the literature and incorporating findings from an earlier qualitative study conducted at the DFCM that focused on leadership issues.12 The questionnaire was pretested, revised, and pilot-tested before implementation. The final questionnaire contained 12 sections and collected information about demographic characteristics; practice patterns; teaching, clinical, administration, and research activities; mentorship received and given; past and present leadership roles; training needs and preferences; job satisfaction; health status; stress; and burnout. The questionnaire also collected information about perceptions of supports provided, recognition, communication, retention, workload, teamwork, respect, resource distribution, remuneration, and infrastructure support. Faculty members were contacted via personalized e-mails that included personalized links to the online survey. Survey implementation followed a modified Dillman approach13 that included several survey promotion activities designed to legitimize the survey, incentives (prizes), and up to 7 personalized contacts. White et al described the design and implementation methods in detail.14 The questionnaire is available from the authors.
Analysis
We analyzed the data using SPSS, version 21. Before analysis, we decided on the most appropriate ways to recode categorical data. Questions that asked participants to select 1 response on a 5-point Likert scale were collapsed into 2 meaningful categories. For example, we collected information about several distinct constructs, such as perceptions of workload, retention, teamwork, respect, and support. Each of these constructs consisted of several questions that were each rated using a 5-point Likert scale (poor to excellent). We created a mean score for each of the constructs and then dichotomized these means into ratings of less than 4 (where 1 was poor, 2 was fair, and 3 was good) and 4 or greater (where 4 was very good and 5 was excellent) to determine an overall rating for each construct. The rationale for this approach was that individual questions within each construct were conceptually correlated, with high internal consistency (ie, values of Cronbach α near unity). The resulting constructs (eg, perceptions of workload, teamwork) were dichotomous variables that represented either ratings of very good and excellent or poor, fair, and good. Similarly, we created 1 overall composite variable for all local department ratings (including support, recognition, communication, leadership, etc). We did not recode questions that employed continuous scales.
The outcome variable was the response to the question, “How would you rate the overall quality of the mentoring that you have received in each of the following areas?” The 6 areas were overall career, clinical practice, teaching, leadership, research, and work-life balance. Respondents ranked each role on the 5-point Likert scale. For bivariate analyses (examining variables associated with receiving high-quality mentorship) and multivariate logistic regression analysis, we created a composite outcome for mentorship quality, comparing ratings of very good and excellent in any of the 6 areas with ratings of good, fair, and poor in all of the 6 roles. Before bivariate analyses, we selected the survey questions associated with receiving very good or excellent mentorship, based on the literature and our own hypotheses.
As appropriate, t tests and
2 tests were used to identify which variables were statistically associated with receiving high-quality mentorship. For each of these variables, we calculated descriptive statistics (numbers and percentages or means and standard deviations) along with P values, odds ratios, and 95% confidence intervals. We used a probability level of less than .05 to determine statistical significance. Variables found to be statistically significant in the bivariate analysis were then fitted using multivariate regression analysis to obtain a parsimonious set of variables that were independently associated with high-quality mentorship. We obtained ethics approval to conduct this study from the University of Toronto Research Ethics Board.
RESULTS
Participant characteristics
The survey was administered to a total of 1029 faculty, 687 of whom responded (66.8% response rate). The mean age of respondents was 47.5 years (range 29 to 82) and women represented 52.3% of the sample (Table 1). By far most respondents had a medical degree (94.5%), with the remainder being doctoral researchers and allied health professionals. Most respondents worked in an interprofessional practice (56.7%), with group practices being the next most popular environment (36.5%), followed by solo practice (6.8%).
Demographic and practice characteristics of respondents
Faculty mentorship experiences
Most respondents received mentorship in at least 1 area (Table 2). Informal mentorship was the most frequently reported type of mentorship received, followed by combined formal and informal mentorship. Most respondents indicated that it was important to receive mentorship for their overall career, clinical practice, teaching, leadership, and work-life balance, whereas less than one-third perceived research mentorship as being important to receive. More than three-quarters of the respondents indicated they provided overall career and clinical mentorship, with smaller percentages providing teaching, leadership, and work-life balance mentorship, and only 25.3% indicated ever providing research mentorship. The overall ratings of high-quality mentorship received (ie, very good and excellent combined) were highest for clinical mentorship, followed by teaching, overall career, leadership, work-life balance, and research mentorship.
Family medicine faculty mentorship experience by type of mentorship
Bivariate analyses
Of the 597 faculty members who rated the overall quality of the mentorship they had received, 354 (59.3%) rated at least 1 area of mentoring as very good or excellent (Table 3). Of the 30 variables found to be significantly associated with overall quality of mentorship on bivariate analysis (Table 3), 12 were related to respondent ratings of their local department (eg, ratings of support, recognition, communication, leadership, teamwork); 1 was related to a rating of infrastructure support; 2 were teaching activities; 1 was related to professional development; 2 were related to leadership; 2 were mentorship variables; 2 were related to burnout; 3 were job satisfaction variables; 2 were health status variables; and 3 were demographic or practice characteristics.
Variables associated with overall rating of mentorship quality: Of the 597 total respondents who rated the overall quality of mentorship, 354 respondents rated at least 1 area as very good or excellent.
Multivariate logistic regression analysis
The final adjusted multivariate logistic regression model identified 5 factors that were independently associated with receiving high-quality mentorship: overall positive faculty ratings of their local department, positive faculty ratings of infrastructure support at their primary practice setting, greater frequency of mentorship, being a junior faculty member, and working in a family practice teaching unit (Table 4).
Final logistic regression model of variables associated with receiving high-quality mentorship (rating of very good or excellent): Hosmer-Lemeshow goodness-of-fit test = 0.26; ρ2 (McFadden pseudo R2) = 0.12 (values between 0.2 and 0.4 suggest a very good fit); N = 557.
DISCUSSION
This study describes the types and characteristics of mentorship activity and the variables that are associated with the perception of high-quality mentorship in a distributed academic family medicine department. The large number of faculty members and the range of practice settings make this relevant for other academic departments aiming to build a culture of mentorship in their organizations.
Several findings add to the existing literature. First, 95.0% of respondents had received mentorship, which compares favourably with other academic medical organizations, where the numbers ranged from 45% to 98%.9,15–17
Second, we found that informal mentoring relationships among faculty predominated. Informal mentoring occurs when mentors and mentees connect naturally without the aid of a third party. This can have implications for the success of the relationship.6 Straus and Sackett maintain that “although successful mentoring can develop through formal assignment, the relationship varies with the individuals involved and will succeed or fail depending on whether a personal connection is in place or can develop over time.”7 During the 5 years before this study, various supports and initiatives were developed in the DFCM to support both formal and informal mentoring relationships.18
Third, nearly 60% of faculty respondents rated at least 1 area of mentorship as very good or excellent, with mentorship in the areas of clinical practice and teaching being rated highest, followed by overall career, leadership, work-life balance, and research mentorship. These results might reflect the reality that most DFCM faculty members are clinician-teachers who work in community sites, where they might not have protected time for research or other scholarly activities. Other studies of mentorship experiences have shown, however, that faculty members with more teaching and patient care responsibilities were less likely to receive mentorship than faculty members with a research focus.17
Fourth, our multivariate analysis identified 5 contextual and individual factors that were independently associated with high mentorship quality. Higher mentorship frequency (monthly or greater) was associated with high quality, suggesting that frequency of contact between mentor and mentee shows commitment to their relationship, which has been reported by others to be important.15,16,19 Context emerged as being important: positive ratings of the local department (including support, recognition, communication, and teamwork) and of infrastructure support (including people and resources), as well as working in a teaching unit, were all independent predictors of perceived high-quality mentorship. These findings reinforce the conclusion of a systematic review of the qualitative literature on the importance of a “facilitating environment” for successful mentorship.3 Faculty in teaching units might perceive mentorship quality to be higher owing to proximity to multiple potential academic mentors or more protected academic time, neither of which was measured in this study. Nevertheless, an important finding for a department that relies on distributed and community-based academic sites is that perception of local department support was the strongest predictor of high ratings for mentorship.
Finally, junior faculty were more likely to rate the quality of the mentorship they had received as very good or excellent; these faculty members might have been encouraged to seek out mentors or take advantage of mentoring opportunities. This finding is in contrast to a similar study by Riley et al, in which junior faculty members were less likely to be satisfied with mentorship in their academic family medicine department.9
The bivariate analysis also revealed that high-quality mentorship was positively associated with multiple potential beneficial outcomes for faculty such as higher levels of job satisfaction, less stress, and more interest in pursuing professional development and leadership opportunities. These findings serve to strengthen the reported benefits of mentorship both for the faculty members involved and the organization.5–7,15
Strengths and limitations
This study has a number of strengths, including the comprehensive survey design, the very good response rate, and the rigorous methodology used for data analysis. Another strength is that this study assessed mentorship as part of a comprehensive survey that addressed a range of personal, environmental, and work-related issues. Compared with surveys that focus solely on mentorship or the effect of a mentorship program, the current study permitted analysis of a broader range of contextual factors that can affect mentorship quality. It is limited by the fact that it was a single-site, cross-sectional survey delivered at a single point in time.
Conclusion
This study showed that 3 contextual factors (positive perceptions of the local department, infrastructure support, and working in a teaching unit), junior faculty status, and frequency of mentorship were associated with perceptions of high mentorship quality for faculty in an academic department of family medicine. Because higher mentorship quality was strongly associated with a supportive context, strategies that strengthen the local culture should enhance mentorship quality. Further, mentorship quality might need to be enhanced for mid-career or senior faculty members. With rising expectations for academic family physicians to successfully engage in clinical and education research and scholarship, high-quality mentorship in these domains will be crucial. Ensuring access to opportunities for mentorship and fostering a culture of mentorship for all academic physicians both within and outside of formal teaching units will become increasingly important.
Acknowledgments
We thank Ms Robyn Butcher, librarian for the Department of Family and Community Medicine, and Ms Nicole S. Ryan for their support and contributions.
Notes
EDITOR’S KEY POINTS
Mentorship is a vital component of faculty development. This study aimed to examine family medicine faculty members’ experiences with mentorship and to identify factors associated with receiving high-quality mentorship.
Most respondents had received and provided mentorship in various areas, and almost 60% had received very good or excellent mentorship in at least 1 area. Informal mentorship was most common. Most indicated that it was important to receive mentorship for their overall career, clinical practice, teaching, leadership, and work-life balance, whereas less than one-third perceived research mentorship to be important.
Higher mentorship quality was strongly associated with a supportive context, so strategies that strengthen the local culture should enhance mentorship quality. Further, mentorship quality might need to be enhanced for midcareer or senior faculty members, as they were less likely to report high-quality mentorship.
POINTS DE REPÈRE DU RÉDACTEUR
Le mentorat est un élément essentiel du développement d’un professeur. Cette étude voulait déterminer l’expérience qu’en ont les professeurs de médecine familiale et identifier les facteurs permettant d’offrir un mentorat de qualité supérieure.
La plupart des répondants avaient eu un mentor et avaient eux-mêmes agi comme mentor dans différents domaines; près de 60 % d’entre eux avaient profité d’un mentorat de très bonne ou d’excellente qualité dans au moins un domaine. Le plus souvent, il s’agissait d’un mentorat informel. La plupart des répondants soulignaient l’importance d’avoir d’un mentor pour l’ensemble de la carrière, pour la pratique clinique, l’enseignement et le leadership, et pour l’équilibre travail-famille, alors que moins d’un tiers estimaient que le mentorat était important pour la recherche.
On a observé une forte corrélation entre la présence d’un soutien du milieu et un mentorat de très bonne qualité, de sorte que les stratégies qui favorisent un tel soutien devraient améliorer la qualité du mentorat. De plus, il y aurait peutêtre lieu d’améliorer la qualité des mentorats offerts par les professeurs plus âgés ou en milieu de carrière, puisque ces derniers étaient moins susceptibles de rapporter un mentorat de grande qualité.
Footnotes
This article has been peer reviewed.
Cet article a fait l’objet d’une révision par des pairs.
Contributors
Drs Stubbs, White, and Kwong conceived the study as members of the Department of Family and Community Medicine Academic Leadership Task Force and the professional development program. All authors contributed to study design. Dr Krueger designed the survey and performed data acquisition, supported by all authors. Mr Meaney and Dr Krueger analyzed the data and all authors contributed to data interpretation. Dr Stubbs wrote the initial draft of the paper and all authors provided critical revision.
Competing interests
None declared
- Copyright© the College of Family Physicians of Canada
References
In this issue
