Factors Influencing the Selection of General Internal Medicine Fellowship Programs A National Survey Enrico Caiola, MD, David Litaker, MD, MSc
Although criteria are available to guide the selection of general internal medicine (GIM) fellowship programs, the factors actually used in this process are unclear. Using a survey of current GIM fellows, we determined that most received information from their residency advisors, and many viewed them as the most important source of fellowship information. Program location was the top selection factor for fellows, followed by research opportunities, availability of a mentor, and the reputation of the program. This information may be useful to both fellowship candidates as an additional selection guide and to program directors seeking to best structure and market their fellowships. KEY WORDS: fellowships and scholarships; career choice; questionnaires; internal medicine. J GEN INTERN MED 2000:15;656–658.
S
electing a general internal medicine (GIM) fellowship program can be an arduous task, requiring consideration of many factors, both personal and professional. The Society of General Internal Medicine (SGIM) has both outlined requirements for fellowship programs1 and developed guidelines2 to help candidates evaluate their quality. Several studies have evaluated factors that influence medical students’ selection of residency programs.3,4 Although similarities may exist with the fellowship selection process, an explicit understanding of these factors may be useful in developing new programs, or redirecting recruitment and marketing of existing fellowships. Furthermore, this information may help candidates better evaluate programs. The purpose of this study was to identify these factors and the relative importance of each in the fellowship selection process.
METHODS A small group of GIM faculty and fellows at our institution developed a 36-item questionnaire addressing sources of information, the importance of perceived program quality and characteristics (using the Directory of
General Internal Medicine Fellowship Programs guidelines2), the interview process, program location, and family considerations in fellowship selection. Questions had 1 of 4 formats: best answer, 5-point Likert scales (1 ⫽ not very important, 5 ⫽ very important in fellowship selection), rank order, and open-ended responses. Fellows were also asked to select one anticipated career choice as either a clinician-researcher, clinician-educator, clinician only, administrator or other position. These categories were provided without a definition. Active, full-time fellowship programs were identified from the 1996 SGIM Directory, and the number of fellows in each program was confirmed by telephone. We were able to survey 30 of the 60 listed programs. Of the 30 programs not surveyed, 8 reported no fellows enrolled in their program; 6 programs no longer existed; 2 programs did not provide their current number of fellows despite several requests (estimated 14 fellows); 8 programs were directed specifically towards medical informatics, cardiovascular disease prevention, cancer epidemiology, pharmacology, or reproductive endocrinology; 2 were critical care or geriatrics programs; 2 were Robert Wood Johnson (RWJ) Clinical Scholars Programs; and 1 was a part-time program. We also excluded our own program since our fellows were involved in the design of the study. We did not survey programs that did not provide their current number of fellows since this would have prevented the determination of an accurate number of total fellows. Programs that focused only on medical informatics or specialty areas of epidemiology were excluded from the study since they were not characteristic of GIM fellowships (i.e., oncology, pharmacology, or reproductive epidemiology). We excluded the RWJ Clinical Scholars Programs because they included fellows from several specialties (e.g., surgery, pediatrics) in addition to internal medicine and have a separate application and fellow selection process. In October 1998, surveys were mailed to 27 fellowship directors (heading 30 programs) for distribution to their current GIM fellows (n ⫽ 146). All surveys were completed anonymously.
RESULTS Received from the Department of General Internal Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio (EC, DL). Presented in part at the 22nd annual meeting, Society of General Internal Medicine, San Francisco, Calif, May 1, 1999. Address correspondence to Dr. Caiola: Department of Medicine, Rochester General Hospital, 1425 Portland Ave., Rochester, NY 14621 (e-mail:
[email protected]). 656
One hundred nine fellows returned the survey for a response rate of 75%, with at least 1 respondent in 83% of the programs. The majority (59%) of fellows anticipated careers as clinician-researchers (i.e., clinician-investigators or physician-investigators); 35% anticipated careers as clinician-educators. Fifty-eight percent of respondents were men; 62% were married.
JGIM
657
Volume 15, September 2000
Responding fellows received information about fellowship programs from several sources including their residency advisor (71%), the SGIM directory (28%), and journal advertisements (4%). Forty-five percent listed their residency advisor or a faculty member as their most important source of fellowship information. The availability of a mentor was important or very important in fellowship program selection for most fellows (85%), as was a positive interview experience (84%), protected time for research or teaching (81%), support for research (78%), national reputation of the program (76%) and access to an advanced degree (70%). Less important factors included a written fellowship curriculum (28%), government sponsorship of the program (17%), and salary (17%). To further elucidate the importance of location and family requirements in the selection process, fellows were asked to rate the importance of the program’s city, its climate, proximity to family, and suitability for a spouse or significant other. Most fellows chose “suitable location for spouse or significant other” (67%) or “interest in the city” (58%) as important or very important in their selection process (Table 1). When asked to identify their top 3 selection factors, 32% ranked location highest. Research opportunities, the availability of a mentor, and national reputation were ranked highest by 16%, 15%, and 14% of the fellows, respectively (Table 2). No significant differences existed by gender, marital status or anticipated career choice as either a clinician-researcher or a clinician-educator, in post
hoc analyses of the top 3 fellowship selection factors. Open-ended responses did not indicate any additional selection factor domains from those present in our survey.
DISCUSSION This study suggests the type and relative importance of factors that current fellows used in selecting their GIM fellowship programs. Several factors, such as location and national reputation, are beyond the immediate control of program directors. However, the availability of mentors, research opportunities, protected time for projects, advanced degrees and the quality of the fellowship interview, are factors that can be enhanced and emphasized. Indeed, these findings are similar in content to the SGIM Directory guidelines2 on GIM fellowship evaluation. Our results parallel findings for fellows in other specialties. A recent national survey of maternal-fetal medicine fellows also identified location, academic reputation and research opportunities as the most important fellowship selection factors.5 Furthermore, this study highlighted the positive effects that mentors have on program satisfaction, completion of a thesis project, and entrance into academic careers.5 An earlier study assessing primary care research fellowships concluded that more attention should be devoted to mentoring.6 These congruent findings suggest that similarities in selecting programs exist across specialties.
Table 1. Rating of Individual Factors Important in General Internal Medicine Fellowship Program Selection
Factor Perceived quality of fellowship program National reputation Placement of graduates Research renown Government sponsorship Robert Wood Johnson sponsorship Characteristics of fellowship program Available mentor Protected time Research support Available advanced degree Accomplishments of previous fellows Teaching opportunities Written curriculum Salary Interpersonal factor Interview experience Location Spouse/significant other Interest in city Near family Climate
Mean Score
Standard Deviation
% of Fellows Rating Factor as Important or Very Important*
4.02 3.75 3.48 2.42 2.04
1.09 1.18 1.23 1.31 1.18
76 63 53 17 11
4.37 4.17 4.13 3.82 3.74 3.61 2.83 2.51
0.84 0.95 1.03 1.36 1.09 1.16 1.24 1.02
85 81 78 70 62 62 28 17
4.33
0.84
84
3.74 3.53 3.09 2.48
1.56 1.28 1.42 1.18
67 58 43 20
*All responses were based on a 5-point Likert scale: 1 ⫽ not very important, 2 ⫽ not important, 3 ⫽ neutral, 4 ⫽ important, 5 ⫽ very important. A minimum of 98 respondents answered for each factor.
658
JGIM
Caiola et al., GIM Fellowship Programs
Table 2. Distribution of Rankings for the Top 3 Factors Influencing General Internal Medicine Fellowship Program Selection
Selection Factor Location Research opportunities National reputation Availability of mentor Interview experience Residency advisor Teaching opportunities Salary
First, %* 32 16 14 15 5 6 6 1
Level of Importance Third, Second, %* %* 7 22 23 18 13 9 6 0
16 17 15 12 15 12 10 4
Total, %† 55 55 52 45 33 27 22 5
*Percentage of fellows ranking factor as the first, second, or third most important factor in their fellowship selection. †Sum of percentages (first through third) for each selection factor. A minimum of 102 respondents answered for each factor.
Most respondents relied heavily on their residency advisors for fellowship information. Surprisingly, less than a third of fellows used the SGIM Directory, the most comprehensive source of GIM fellowship information currently available. Similarly, journal advertisements did not appear to influence fellowship selection. Our study has several limitations and potential biases. An inability to include all GIM fellowship programs and include all fellows in the United States may have resulted in potentially significant selection and response biases. The exclusion of programs concentrating in medical informatics and specialty areas of epidemiology may, for example, have introduced selection bias. In addition, inclusion in the SGIM Fellowship Directory, the sole source for information on GIM fellowship programs used in this study, is voluntary; its listing of fellowships may have been incomplete. When updated in January 1999, the SGIM Fellowship Directory (available online at www.sgim.org) contained only 1 new program not included in this survey. A survey of all GIM fellowship applicants (including those who initially applied for a fellowship but did not pursue training or those who were not accepted in a program) may have been more instructive by identifying factors more generalizable or relevant to the fellowship selection process. By surveying only current GIM fellows, this study may have missed factors or concerns important to applicants who decided not to pursue fellowship training. Logistical constraints, such as obtaining accurate mailing lists of all GIM fellowship applicants and consent for their release, made a complete survey of this type beyond the resources of this project.
Our instrument may also have missed important selection factor domains because it was developed intuitively and was not formally validated. However, qualitative analysis of responses to open-ended questions suggested good content validity. Despite these potential limitations, we have described important factors in fellowship selection for the majority of current GIM fellows and these factors closely coincide with selection factors of fellows from other specialties. This survey was designed to evaluate fellows’ perceptions of factors that influenced their selection of GIM fellowship programs. We did not objectively measure or vary factors, such as the interview process, the quantity or quality of mentoring, the range of available research or teaching opportunities, or the qualities of applicants, in order to establish causal relationships between program characteristics and recruitment results. Also we did not seek to address fellows’ satisfaction with their programs after acceptance. Future research in these areas may be useful in assessing the overall quality of GIM fellowship programs. In summary, location, applicants’ perceptions of research opportunities, the availability of mentors, and a program’s national reputation influence applicants’ selection of GIM fellowship programs. Fellowship program directors can explicitly emphasize these factors to more effectively attract fellows. Finally, fellowship directors might consider the value of directly providing information about their programs to residency program faculty, because this group appears to strongly influence candidates’ selection processes. The authors wish to thank Dr. Mariana Hewson for her assistance with the study design and Jessica Ancker for her careful review of the manuscript.
REFERENCES 1. Society of General Internal Medicine. Policy statement for general internal medicine fellowships. J Gen Intern Med. 1994;9:513–6. 2. Directory of General Internal Medicine Fellowship Programs. Washington, DC: Society of General Internal Medicine; 1996. 3. Flynn TC, Garrity MS, Berkowitz LR. What do applicants look for when selecting internal medicine residency programs? A comparison of rating scale and open-ended responses. J Gen Intern Med. 1993;8:249–54. 4. Simmonds AC, Robbins JM, Brinker MR, Rice JC, Kerstein MD. Factors important to students in selecting a residency program. Acad Med. 1990;65:640–3. 5. Sciscione AC, Colmorgen GH, D’Alton ME. Factors affecting fellowship satisfaction, thesis completion, and career direction among maternal-fetal medicine fellows. Obstet Gynecol. 1998; 91:1023–6. 6. Curtis P, Reid A, Newton W. The primary care research fellowship: an early assessment. Fam Med. 1992;24:586–90.
