Keywords
Retention - Training location - Internal hiring - Residents and fellows - Fellowship - Ophthalmology faculty
Throughout graduate and postgraduate education, physician trainees may want to determine what impact, if any, their choice of training institution will have on where they practice in the future. Indeed, the American Association of Medical Colleges (AAMC) reported that over half (55.5%) of physicians who graduated from residency between 2010 and 2019 practice in their state of residency training.[1] In addition, several studies have sought to characterize trainee retention in various medical specialties on an institutional level. One such study in the field of radiation oncology found that 40.9% of graduates working in academics were practicing at the same institution as where they completed residency.[2] Similarly, a study of academic plastic surgeons who graduated within the past 3 years found that 38.6% of recent graduates practice at their institution of residency or fellowship training.[3]
When examining retention, other important considerations include whether specific factors influence a trainee's likelihood to practice at their location of residency or fellowship. A few studies have sought to characterize trends in retention by gender. For example, the AAMC found a higher percentage of female physicians relative to male physicians practicing in their state of residency training.[1] A similar difference was also found in radiation oncology, where a greater proportion of women (85.1%) worked in a geographic region where they had previously trained as compared to men (68.4%).[2]
Although some have examined retention in other specialties, there are limited data specific to retention within the field of ophthalmology. In its report on the different medical specialties, the AAMC found that geographic retention rates within ophthalmology were lower than those in many other specialties, with only 36.4% of postgraduate ophthalmologists currently practicing in their state of postgraduate education.[4] Other research has sought to characterize the impact of training location on the leadership role within ophthalmology, finding that many leaders trained at the top academic institutions for ophthalmology.[5]
[6]
[7] However, no study has examined rates of internal hiring of faculty after completion of residency or fellowship at the same institution in the field of ophthalmology.
The purpose of the present study was to use publicly available information on faculty ophthalmologists to determine the percentage of faculty that completed residency or fellowship training at the same institution and what factors are associated with being an internal hire.
Methods
Study Sample
This study was qualified as exempt by the Johns Hopkins School of Medicine Institutional Review Board and was conducted in accordance with the Strengthening the Reporting of Observational Studies in Ophthalmology guidelines for cross-sectional studies. This cross-sectional analysis included 1,246 clinical ophthalmology faculty at the 13 top-ranked institutions listed in the 2021 U.S. News and World Report.[8] The top 13 institutions were chosen to be included because they are the institutions that are ranked in the U.S. News and World Report; there is no definite ranking for other institutions (other top hospitals may be listed as “High Performing in Ophthalmology,” but are not formally ranked). Institutions that were included were the Bascom Palmer Eye Institute (University of Miami Hospital and Clinics), Wills Eye Hospital (Thomas Jefferson University Hospitals), Wilmer Eye Institute (Johns Hopkins Hospital), Mass Eye and Ear (Massachusetts General Hospital), Stein and Doheny Eye Institutes (UCLA Medical Center), Duke University Hospital, University of Iowa Hospitals and Clinics, Kellogg Eye Center (University of Michigan), UCSF Medical Center, Cole Eye Institute (Cleveland Clinic), John A. Moran Eye Center (University of Utah Hospitals and Clinics), New York Eye and Ear Infirmary (Mount Sinai), and Roski Eye Institute (USC). Faculty included were those listed on the websites of the top 13 institutions as ophthalmologists. Primary faculty, emeritus faculty, adjunct, and affiliate faculty were all included. Physicians who were still in training (residents and fellows), faculty who did not complete any residency or fellowship training in ophthalmology, and faculty without a medical doctorate, such as optometrists and research faculty, were excluded. Data collection was performed between June 15 and July 31, 2021.
Data Collection
Data were collected using institutional websites and other online faculty profiles, including LinkedIn, public websites (including Doximity, U.S. News, and Healthgrades), and private practice websites. Gender, graduation year, institution, and location of medical school, residency, and fellowship training were collected. Data were also gathered on current academic rank, leadership roles (department chair, chief of service, fellowship director, residency director, or other leadership such as clinical laboratory directors or medical education deans, for example), subspecializations, and advanced degrees. Subspecialization categories included anterior segment/cornea/refractive surgery, glaucoma, oculoplastics, medical retina, vitreoretinal surgery, neuro-ophthalmology, pediatrics ophthalmology, ophthalmic oncology, ocular immunology/uveitis, and ophthalmic pathology. Additional data, including the number of first and last author publications and lifetime National Institutes of Health (NIH) funding, were collected using PubMed and the NIH RePorter tool. Department size was calculated using the number of ophthalmology faculty for whom data were collected at each institution. The geographic region of each program was determined based on U.S. Census Bureau-designated regions of the United States. City sizes were determined using the US Census Reporter website.[9]
Statistical Analysis
Internal hires were defined as faculty who had completed residency and/or fellowship training at their current institution. External hires were defined as faculty who had completed neither residency nor fellowship at their current institution. Recent graduates were defined as faculty who had completed residency within 5 years of the time of data collection in 2021. For some analyses, internal hires were subdivided into faculty who were internally trained for residency and faculty who were internally trained for fellowship. Statistical analyses were conducted using t-tests or Mann–Whitney tests, chi-squared or Fisher's exact tests, and multivariable logistic regression. To evaluate the risk of being an internal hire, multivariable logistic regression was adjusted for gender, medical school region, years since residency graduation, advanced degrees, current institution size, current city size, NIH funding, and number of publications. Odds ratios (OR) and 95% confidence intervals (CI) were reported.
Results
Demographic Information
Demographic data were collected on 1,246 clinical ophthalmology faculty ([Table 1]). Of the 1,246 faculty, 589 (47.3%) were internal hires from residency or fellowship and 657 (52.7%) were external hires. Overall, the majority of faculty were male (64.6% of external hires and 62.1% of internal hires). There was no significant difference in gender ratios between groups (p = 0.36). Most faculty were currently practicing at institutions in the Northeast (50.7% of external hires and 48.7% of internal hires). The majority of faculty also attended medical school in the Northeast (46.4% of internal hires and 52.1% of external hires). Internal hires were more likely to practice in comparatively smaller departments (78 vs. 121 faculty, p = 0.002), have graduated residency more recently (20 vs. 23, p < 0.001), and were more likely to have obtained advanced degrees in addition to a medical doctorate, including a PhD or a non-PhD advanced degree (19.4 vs. 13.4%, p = 0.015). Internal hires were also more likely to have obtained NIH funding (18.9 vs. 12.6%, p = 0.002), had a greater median number of publications (7 vs. 3, p <0.001), and were more frequently trained in a subspecialty of ophthalmology (84.7 vs. 76.3%, p < 0.001). In an analysis of only recent graduates, there were no significant differences between internal hires and external hires ([Supplementary Table S1]).
Table 1
Demographic information of clinical ophthalmology faculty at 13 top-ranked institutions
|
Factor
|
Level
|
External hires
|
Internal hires
|
p-Value
|
|
N
|
|
657
|
589
|
|
|
Gender
|
Male
|
424 (64.6%)
|
366 (62.1%)
|
0.36
|
|
Female
|
232 (35.4%)
|
223 (37.9%)
|
|
|
Current institution region
|
Northeast
|
332 (50.7%)
|
287 (48.7%)
|
<0.001
|
|
South
|
83 (12.7%)
|
131 (22.2%)
|
|
|
Midwest
|
102 (15.6%)
|
60 (10.2%)
|
|
|
West
|
138 (21.1%)
|
111 (18.8%)
|
|
|
Medical school region
|
Northeast
|
337 (52.1%)
|
273 (46.4%)
|
0.052
|
|
South
|
98 (15.1%)
|
114 (19.4%)
|
|
|
Midwest
|
97 (15.0%)
|
82 (13.9%)
|
|
|
West
|
47 (7.3%)
|
62 (10.5%)
|
|
|
Outside United States
|
68 (10.5%)
|
57 (9.7%)
|
|
|
Year since residency graduation in 2021, mean (SD)
|
|
23 (13)
|
20 (13)
|
<0.001
|
|
Advanced degrees
|
No other advanced degree
|
569 (86.6%)
|
475 (80.6%)
|
0.015
|
|
PhD
|
39 (5.9%)
|
55 (9.3%)
|
|
|
Non-PhD advanced degrees
|
49 (7.5%)
|
59 (10.0%)
|
|
|
No specialization
|
No
|
501 (76.3%)
|
499 (84.7%)
|
<0.001
|
|
Yes
|
156 (23.7%)
|
90 (15.3%)
|
|
|
Anterior segment/cornea/refractive surgery
|
No
|
537 (81.7%)
|
465 (78.9%)
|
0.22
|
|
Yes
|
120 (18.3%)
|
124 (21.1%)
|
|
|
Glaucoma
|
No
|
582 (88.6%)
|
484 (82.2%)
|
0.002
|
|
Yes
|
75 (11.4%)
|
105 (17.8%)
|
|
|
Oculoplastics
|
No
|
614 (93.5%)
|
541 (91.9%)
|
0.33
|
|
Yes
|
43 (6.5%)
|
48 (8.1%)
|
|
|
Medical retina
|
No
|
618 (94.1%)
|
543 (92.2%)
|
0.22
|
|
Yes
|
39 (5.9%)
|
46 (7.8%)
|
|
|
Vitreoretinal surgery
|
No
|
569 (86.6%)
|
508 (86.2%)
|
0.87
|
|
Yes
|
88 (13.4%)
|
81 (13.8%)
|
|
|
Neuroophthalmology
|
No
|
620 (94.4%)
|
561 (95.2%)
|
0.52
|
|
Yes
|
37 (5.6%)
|
28 (4.8%)
|
|
|
Pediatric ophthalmology
|
No
|
604 (91.9%)
|
533 (90.5%)
|
0.42
|
|
Yes
|
53 (8.1%)
|
56 (9.5%)
|
|
|
Ophthalmic oncology
|
No
|
646 (98.3%)
|
579 (98.3%)
|
1.00
|
|
Yes
|
11 (1.7%)
|
10 (1.7%)
|
|
|
Ocular immunology/uveitis
|
No
|
639 (97.3%)
|
568 (96.4%)
|
0.42
|
|
Yes
|
18 (2.7%)
|
21 (3.6%)
|
|
|
Ophthalmic pathology
|
No
|
648 (98.6%)
|
580 (98.5%)
|
0.82
|
|
Yes
|
9 (1.4%)
|
9 (1.5%)
|
|
|
Other specialization
|
No
|
651 (99.1%)
|
581 (98.6%)
|
0.59
|
|
Yes
|
6 (0.9%)
|
8 (1.4%)
|
|
|
Current department size, median (IQR)
|
|
121 (63, 360)
|
78 (63, 185)
|
0.005
|
|
City size (thousands), median (IQR)
|
|
1584 (381, 8337)
|
882 (468, 3980)
|
0.23
|
|
NIH funding
|
No
|
568 (87.4%)
|
477 (81.1%)
|
0.002
|
|
Yes
|
82 (12.6%)
|
111 (18.9%)
|
|
|
Past NIH funding
|
No
|
6 (7.3%)
|
7 (6.3%)
|
0.78
|
|
Yes
|
76 (92.7%)
|
104 (93.7%)
|
|
|
Past R01 funding
|
No
|
33 (43.4%)
|
61 (58.7%)
|
0.043
|
|
Yes
|
43 (56.6%)
|
43 (41.3%)
|
|
|
Active NIH funding
|
No
|
45 (54.9%)
|
54 (48.6%)
|
0.39
|
|
Yes
|
37 (45.1%)
|
57 (51.4%)
|
|
|
Active R01 funding
|
No
|
13 (35%)
|
27 (47%)
|
0.24
|
|
Yes
|
24 (65%)
|
30 (53%)
|
|
|
Publications, median (IQR)
|
|
3 (0, 16)
|
7 (1, 24)
|
<0.001
|
Abbreviations: IQR, interquartile range; NIH, National Institutes of Health; R01, Research Project Grant; SD, standard deviation.
Note: Statistical significance indicated in bold.
Retention by Institution
A median percentage of 53.8% of faculty across all 13 institutions were internal hires. Bascom Palmer had the highest percentage of internal hires at 73.0%, while the Cole Eye Institute had the lowest percentage of internal hires at 20.3% ([Table 2]). The median proportion of faculty who had completed a residency at their current institution (internal residency training) was 28.6% (range: 11.9 to 40.0%). The median proportion of faculty with internal fellowship training was 38.9% (range: 11.6 to 63.9%). Among faculty who had completed external residency training, a median of 27.7% (range: 7.5 to 45.8%) completed their residency and a median of 56.0% (range: 24.6 to 69.2%) completed their fellowship at one of the other top 13 programs.
Table 2
Internal residency and fellowship training among current ophthalmology faculty, separated by current institution
|
Bascom Palmer
|
Wills
|
Wilmer
|
Mass. Eye and Ear
|
UCLA
|
Iowa
|
Duke
|
Michigan
|
UCSF
|
Cole Eye Institute
|
Mount Sinai
|
USC
|
Utah
|
|
N
|
74
|
185
|
78
|
76
|
61
|
39
|
61
|
63
|
37
|
59
|
360
|
32
|
121
|
|
Total internal hires
|
54 (73%)
|
100 (54%)
|
42 (54%)
|
47 (62%)
|
38 (62%)
|
20 (51%)
|
34 (56%)
|
29 (46%)
|
21 (57%)
|
12 (20%)
|
140 (39%)
|
13 (41%)
|
39 (32%)
|
|
Retention from residency
|
|
External residency training
|
55 (74%)
|
111 (60%)
|
53 (68%)
|
48 (63%)
|
45 (74%)
|
25 (66%)
|
47 (77%)
|
45 (71%)
|
24 (65%)
|
52 (88%)
|
254 (71%)
|
26 (84%)
|
82 (78%)
|
|
External at top 13 institutions
|
15 (27%)
|
21 (19%)
|
12 (23%)
|
16 (33%)
|
10 (22%)
|
8 (32%)
|
13 (28%)
|
16 (36%)
|
11 (46%)
|
12 (23%)
|
19 (8%)
|
10 (39%)
|
27 (33%)
|
|
Internal residency training
|
19 (26%)
|
74 (40%)
|
25 (32%)
|
28 (37%)
|
16 (26%)
|
13 (34%)
|
14 (23%)
|
18 (29%)
|
13 (35%)
|
7 (11.9%)
|
102 (29%)
|
5 (16%)
|
23 (22%)
|
|
Retention from fellowship
|
|
External fellowship training
|
26 (36%)
|
65 (56%)
|
43 (61%)
|
31 (44%)
|
29 (50%)
|
23 (64%)
|
29 (51%)
|
33 (61%)
|
25 (69%)
|
38 (88%)
|
199 (77%)
|
19 (66%)
|
41 (65%)
|
|
External at top 13 institutions
|
18 (69%)
|
26 (40%)
|
27 (63%)
|
20 (65%)
|
16 (55%)
|
14 (61%)
|
16 (55%)
|
19 (58%)
|
14 (56%)
|
21 (55%)
|
49 (25%)
|
11 (58%)
|
15 (37%)
|
|
Internal fellowship training
|
46 (64%)
|
51 (44%)
|
28 (39%)
|
40 (56%)
|
29 (50%)
|
13 (36%)
|
28 (49%)
|
21 (39%)
|
11 (31%)
|
5 (12%)
|
59 (23%)
|
10 (35%)
|
22 (35%)
|
Multivariable Regression
Multivariable logistic regression evaluating the risk of being an internal hire was performed and adjusted for gender, medical school region, years since residency graduation, advanced degrees, current institution size, current city size, NIH funding, and number of publications ([Table 3]). A total of 1,210 faculty had complete information and were included in the multivariable analysis. Faculty currently practicing in the Midwest or West were less likely to be internal hires than faculty currently practicing in the Northeast (reference: Northeast; adjusted OR, 0.29; 95% CI, 0.18–0.48; and adjusted OR, 0.49; 95% CI, 0.30–0.78, respectively). Years since completion of residency were a significant predictor of internal hiring, with a greater number of years since graduation being associated with a lower likelihood of internal hiring (adjusted OR, 0.98; 95% CI, 0.97–0.99). Faculty practicing in larger departments were also less likely to be internal hires (adjusted OR, 1.00; 95% CI, 0.99–1.00). Further, faculty with NIH funding other than an R01 grant were more likely to be internal hires (reference: no funding, adjusted OR, 1.83; 95% CI, 1.12–2.96).
Table 3
Multivariable logistic regression for internal hiring of ophthalmology faculty
|
Odds ratio
|
95% CI
|
p-Value
|
|
Gender
|
|
Female vs. male
|
0.97
|
0.76–1.25
|
0.836
|
|
Current institution region
|
|
|
|
|
Northeast
|
Reference
|
|
|
|
South
|
0.80
|
0.51–1.25
|
0.331
|
|
Midwest
|
0.29
|
0.18–0.48
|
<0.001
|
|
West
|
0.49
|
0.31–0.78
|
0.003
|
|
Years since completion of residency
|
0.98
|
0.97–0.99
|
<0.001
|
|
Advanced degrees
|
|
No advanced degree
|
Reference
|
|
|
|
PhD
|
1.32
|
0.78–2.22
|
0.298
|
|
Non-PhD advanced degrees
|
1.05
|
0.67–1.62
|
0.84
|
|
Current department size
|
1.00
|
0.99–1.00
|
0.018
|
|
City size (thousands)
|
1.00
|
1.00–1.00
|
0.575
|
|
NIH funding
|
|
No
|
Reference
|
|
|
|
Yes, no R01
|
1.82
|
1.12–2.96
|
0.015
|
|
Yes, R01
|
0.90
|
0.53−1.53
|
0.706
|
|
Publications
|
1.00
|
1.00−1.01
|
0.112
|
Abbreviations: CI, confidence interval; NIH, National Institutes of Health; R01, Research Project Grant (R01).
Impact of Leadership Roles on Internal Hiring
Of the 589 internal hires across all 13 institutions, 166 (29.1%) had a leadership position, whereas 137 (22.7%) of external hires had a leadership position (p =0.012, [Table 4]). Subdividing into leadership roles including chair or chief, fellowship director, or residency director did not yield significant differences between internal and external hires.
Table 4
Retention of ophthalmology faculty by leadership roles
|
Factor
|
Level
|
External hires
|
Internal hires
|
p-Value
|
|
N
|
|
657
|
589
|
|
|
Overall leadership
|
No
|
467 (77.3%)
|
405 (70.9%)
|
0.012
|
|
Yes
|
137 (22.7%)
|
166 (29.1%)
|
|
|
Chair or chief
|
No
|
605 (92.1%)
|
529 (89.8%)
|
0.16
|
|
Yes
|
52 (7.9%)
|
60 (10.2%)
|
|
|
Fellowship director
|
No
|
614 (93.5%)
|
534 (90.7%)
|
0.067
|
|
Yes
|
43 (6.5%)
|
55 (9.3%)
|
|
|
Residency director
|
No
|
652 (99.2%)
|
580 (98.5%)
|
0.20
|
|
Yes
|
5 (0.8%)
|
9 (1.5%)
|
|
Discussion
Our study found that, among the top 13 institutions in ophthalmology, the location of residency or fellowship training in ophthalmology is key to the future location of practice. Indeed, among the 1,246 clinical ophthalmology faculty included in our cohort, 47.3% were internal hires from residency or fellowship. Several factors modulated internal hiring, including department size, subspecialty training, and number of publications. Further, the number of years since residency graduation was a significant predictor of internal hiring, although our study likely was not powered enough to detect significant differences in hiring practices among recent graduates. Interestingly, even for those faculty who were not internal hires from residency or fellowship, a sizeable proportion trained at one of the other top 13 institutions for residency (27.7%) or fellowship (56.0%). Together, our results underscore the strong influence of training institutions on the future location of practice within ophthalmology along with the key role of interinstitutional networks formed by trainees at top programs. Notably, we found that the location of fellowship training had a greater effect than the location of residency training. Applicants to fellowship programs may want to closely consider their desired location of future practice.
Our study found several differences when compared to similar studies both within and outside of ophthalmology. For example, others have reported that gender influences internal hiring, with women being more likely to remain in their geographic area of training.[1] However, our study found no difference between the proportion of women who were internal hires and those who were not. Of note, we did find a greater proportion of male faculty relative to female faculty, a disparity that has been previously documented in the field of ophthalmology.[5]
[10] Prior studies have found that gender modulates both intent to pursue subspecialty training and pursuit of specific subspecialties within ophthalmology; however, such an analysis was not within the scope of our present study.[11]
[12] In addition, a previous study examining retention within plastic surgery demonstrated that recent graduates may be less likely to be internal hires compared to faculty who did not graduate as recently.[3] In our study, we found that, on the contrary, fewer years since completion of residency was a significant predictor of internal hiring. We hypothesize that this may be due to the desire to remain in each location for a given amount of time as well as the ties that are created to a particular institution being strongest immediately after training.
We also report on several novel findings. Our study found that internal hires were more frequently on faculty at institutions with comparatively smaller department sizes. While the programs we studied are generally larger than most other ophthalmology programs in the United States, the correlation between size and internal hiring may be attributable to stronger faculty connections in smaller departments and potential variations in mentorship and research approaches across departments. Future studies will be needed to determine whether this trend is seen across a wider range of institution sizes. Faculty–trainee connections may also play a role in the greater percentage of internal hires with leadership positions.
Our study has several limitations. First, publicly available online data may not be the most accurate or up-to-date information, and some pertinent information such as training institution may not be listed online, leading to the exclusion of some faculty from multivariable analyses. Because we did not contact faculty as part of this study, we were unable to report on associations between self-identified race and retention. Further studies will need to be conducted to determine the extent to which race may impact internal hiring, as this would be important to the critical need of increasing the diversity of the ophthalmology workforce. Moreover, the use of only publicly available online data limits the types of collectable data; personal information such as marital status, parenthood status, familial ties, where individuals were raised, duration of hire, and research interests, while certainly important to individual retention, is impossible to collect using our methodology. These factors will be important to analyze in future studies. Given our methodology, we are also unable to discern whether trends in internal retention are driven primarily by trainee preference or by institutional selection (e.g., some programs may preferentially accept trainees that are most likely to remain on faculty). Future studies are warranted to further assess this relationship. In addition, limiting our analysis to the top 13 institutions may lead to results that are not generalizable across all academic ophthalmology programs. Indeed, our geographical data are difficult to generalize given that institutions included in our analysis were predetermined by rankings and that many top-ranked ophthalmology programs are generally large programs and, for the most part, located in the Northeast. Even the comparatively smaller institutions in our study are larger than most ophthalmology programs; results regarding program size may thus not be generalizable across other institutions. The study's cross-sectional nature also limits interpretation given that it does not account for movement from institution to institution over the course of a career and does not allow for determination of length of employment at a given institution.
Future directions for our work include performing longitudinal studies to capture internal hiring trends throughout a career in ophthalmology. Additionally, expanding our dataset to include more institutions would generate results that are applicable to all ophthalmology trainees, whether or not their goal is to practice at a large academic center. Finally, additional studies may help characterize internal hiring rates across a broader range of specialties.
Conclusions
These findings have several implications for current and future trainees in the field of ophthalmology and, more broadly, for trainees across medical fields. Trainees should be aware that their choice of training institution for residency or fellowship may have a strong impact on where they practice in the future. Accordingly, trainees may want to consider their longer-term goals when selecting a training program.
