Abstract
Objective:
To describe trends in the surgical and medical retina fellowship match and applicant characteristics associated with matching in retina fellowship.
Design:
Retrospective study.
Subjects:
Ophthalmology fellowship applicants who applied through the San Francisco (SF) Match.
Methods:
Publicly available SF Match data were used to describe trends in the number of programs participating and positions offered, filled, and left vacant in the retina fellowship match from 2014 to 2019. De-identified applicant data for match cycles 2010–2017 were stratified by match status, and characteristics were compared across groups. Trends in matched applicant characteristics were evaluated using a linear regression on log-transformed variables. A multivariable logistic regression was used to determine applicant characteristics that were associated with a successful match
Main Outcome Measures:
Match status.
Results:
From 2014 to 2019, the number of programs participating, positions filled, and positions left vacant in the retina fellowship match increased from 101 to 119 (p=0.010), 136 to 160 (p=0.005), and 18 to 37 (p=0.045), respectively. Compared with unmatched applicants, matched applicants were more likely to have graduated from a top 10 residency program, U.S. residency or medical school; hold a U.S. visa (J-1, H-1B, or O1); distribute more applications; complete more interviews; rank more programs; and score higher on USMLE step exams 1–3. Matched applicants completed a median of 10 interviews. After controlling for potential covariates, graduating from a U.S. residency (OR: 2.08, CI: [1.48, 2.92]), a top 10 residency (OR: CI: 1.74, [1.07, 2.84]), having an allopathic medical degree (MD, OR: 2.39, CI: [1.08 5.33]), completing more interviews (OR: 1.28, CI: [1.23, 1.33]), and scoring higher on USMLE Step 3 (OR: 1.01, CI: [1.01, 1.03]) were associated with matching into retina fellowship.
Conclusions:
Although the number of programs participating and positions offered in the retina fellowship match are increasing, the number of positions filled remained relatively stagnant. Factors associated with matching in both medical and surgical retina included graduating from a U.S. and top 10-ranked residency program, having an MD, completing more interviews, and scoring higher on USMLE Step 3.
Keywords: retina fellowship, medical education, ophthalmology
Précis
The number of retina fellows increased. Factors associated with matching into retina fellowship included graduating from a U.S. or top 10 residency, having an MD, completing more interviews, and scoring higher on USMLE Step 3.
Introduction
The proportion of ophthalmology residents pursuing fellowship training has gradually increased over the past 30 years, from 43% in 1987 to 72% in 2018.1,2 Retina fellowship has consistently offered the most positions and matched the greatest number of fellowship applicants since 2002,3 with an average of 35.5% of successful ophthalmology fellowship applicants matching into retina during the 2010 through 2017 match cycles.1 “Retina” fellowship is an umbrella term for two different paths: medical retina, which is typically a one-year fellowship, and surgical retina, which is currently a two-year fellowship.
Retina fellowship – particularly surgical retina – is considered one of the most competitive ophthalmology subspecialties. In 2019, vitreoretinal fellowship applicants submitted nearly twice as many applications as the average fellowship applicant.4 In addition, from 2010 to 2017, there was a statistically significant decrease in the retina match rate, perhaps due to a greater increase in the number of applicants compared with the number of available positions.1 Unfortunately, it is challenging to accurately interpret reports regarding the retina fellowship match, because medical and surgical retina fellowships are typically grouped.
Several studies have described factors influencing post-residency career decisions among ophthalmology residents.2,5,6 Factors associated with the decision to pursue fellowship training were the desire to acquire new skills and work with newer, more specialized technology.5 Applicants also cited mentor relationships and specific clinical rotations as influential for their career pathway.6 For those pursuing fellowship training, it is helpful to understand which factors are weighted most highly by programs when selecting between applicants. Survey-based studies of applicants and program directors at various ophthalmology fellowship programs showed that interviews, letters of recommendation from subspecialty faculty, and communication skills were the most important factors aiding the decision process.4,7 Applicant performance in medical school, including United States Medical Licensing Exam (USMLE) step scores, Ophthalmic Knowledge Assessment Program (OKAP) scores, and the personal statement were valued less highly.4 Compared with cornea and glaucoma fellowship program directors, however, retina fellowship program directors placed more value on the quality of an applicant’s residency program and OKAP scores.7
Recent studies using applicant data from the SF Match found that graduating from a U.S. residency program and ranking more programs were associated with increased likelihood of matching into an ophthalmology fellowship program.1 Similarly, graduating from a U.S. residency program and completing more interviews were associated with increased likelihood of matching into glaucoma fellowship.8 No such study has evaluated objective applicant characteristics associated with matching into retina fellowship, specifically. Understanding successful retina fellowship applicant characteristics may be of interest to prospective applicants, program directors, and workforce planning committees.
Materials and Methods
The present study was determined to be exempt by the Johns Hopkins University School of Medicine Institutional Review Board. This study adhered to the tenants of the Declaration of Helsinki.
Data Source
Individual, de-identified applicant data were provided by the Association of University Professors of Ophthalmology (AUPO) for the 2010 to 2017 San Francisco (SF) match cycles. The following variables were used for analysis: match year; whether the applicant’s graduating residency program ranked in the top 10 programs at the time of the match according to U.S. News & World Report best hospitals in Ophthalmology; graduating residency location (within or outside the United States); graduating medical school location (in the U.S., Canada, or international); visa status (U.S., Canada, or international); if the applicant graduated from an allopathic medical school; number of applications distributed, interviews completed, and programs ranked; and USMLE exam scores for step 1, 2, and 3. Publicly available, conglomerate data were gathered from the SF Match website for the 2014 to 2019 retina fellowship match cycles. Variables used for analysis included the number of programs participating in the retina fellowship match, number of positions offered and filled, and number of positions left vacant.
Statistical Analysis
Trends in the number of programs participating in the retina fellowship match, the number of positions offered, the number of positions filled, and the number of vacancies were determined for match cycles 2014 to 2019. Percent change per year was calculated using linear regressions on log-transformed response variables.
Individual applicant data were stratified by match status (matched in retina versus did not match in any sub-specialty) for the 2010 to 2017 match cycles. We were unable to differentiate between successful applicants who matched in medical versus surgical retina or distinguish between unmatched applicants by the sub-specialties to which they applied. Matched and unmatched applicant characteristics were compared using Pearson’s chi-squared test for categorical variables and Wilcoxon rank-sum test for continuous variables. Yearly variations in matched retina applicant characteristics were assessed from 2010 to 2017. Changes were evaluated for statistical significance using Pearson’s chi-squared test for categorical variables and Kruskal-Wallis test for continuous variables.
Finally, a multivariable logistic regression model was used to determine which characteristics were associated with a successful match after controlling for potential covariates. The following variables were used in the model: location of graduating residency program (inside or outside of the US), ophthalmology residency program rank (within or out of the top 10), medical degree (allopathic or not), the number of applications submitted, the number of interviews completed, and scores from USMLE step exams 1, 2, and 3. All applicant characteristics were evaluated for correlation in a pairwise fashion. A Pearson’s correlation coefficient of magnitude ≥ 0.7 was considered to suggest that variables were correlated. The number of programs ranked was not included in the regression analysis due to correlation with the number of interviews completed. Due to correlation with residency program location, applicant visa status and medical school location were not included in the regression analyses. Observations with missing data for the variables included in the logistic regression were excluded from all analyses. All statistical analyses were completed using RStudio (R version 4.0.2) with statistical significance set at a p-value of p <0.05.
Results
For this paper, “retina fellowship” encompasses both medical and surgical retina fellowships. Over the 2014 to 2019 match cycles, the number of programs participating in the retina fellowship match increased from 101 to 119 programs (on average 2.8% per year, p=0.010, Table 1). The number of positions offered, filled, and left vacant increased from 136 to 160 positions (on average 2.6% per year, p = 0.005), 118 to 123 (on average 0.9% per year, p=0.078), and 18 to 37 (on average 11.5% per year, p = 0.045), respectively.
Table 1.
Retina Fellowship Match Trends, 2014–2019
| 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | Overall % change | % Change per yeara | 95% confidence intervala | p-valuea | |
|---|---|---|---|---|---|---|---|---|---|---|
| Number of Participating Programs | 101 | 103 | 107 | 111 | 108 | 119 | 18% | 2.8 | 1.6, 4.1 | 0.010 |
| Number of Positions Offered | 136 | 143 | 145 | 148 | 148 | 160 | 18% | 2.6 | 1.7, 3.6 | 0.005 |
| Number of Positions Filled | 118 | 118 | 124 | 123 | 123 | 123 | 4% | 0.9 | 0.1, 1.6 | 0.078 |
| Number of Vacancies | 18 | 25 | 21 | 31 | 25 | 37 | 106% | 11.5 | 3.5, 20.2 | 0.045 |
percent change and p-values calculated using linear regression models on log-transformed response variables, where % change = 100*(exp(coefficient)-1)
Compared with applicants who did not match in any sub-specialty during the 2010 to 2017 match cycles, applicants who matched into retina fellowship were more likely to have graduated from a top 10 residency program (14.1% of matched versus 4.5% of unmatched applicants, respectively, p<0.001, Table 2), U.S. residency program (87.3% vs. 42.2%, p<0.001), and U.S. medical school (82.6% vs. 39.1%, p<0.001), and hold a U.S. visa (81.9% vs. 37.2%, p<0.001). Matched applicants were less likely to have graduated from an international medical school (13.3% vs. 57.2%, p<0.001). Most applicants had allopathic medical degrees (98.0% of matched and 96.6% of unmatched applicants, respectively), and there was no significant difference between groups (p = 0.145).
Table 2.
Applicant Characteristics Stratified by Applicant Match Status, 2010–2017
| Matched in Retina | Total Unmatched | p-valuea | ||
|---|---|---|---|---|
| N=739+ | N=801 | |||
| Match Year, n (%) | 2010 | 90 (12.2) | 117 (14.6) | 0.338 |
| 2011 | 98 (13.3) | 110 (13.7) | ||
| 2012 | 96 (13.0) | 117 (14.6) | ||
| 2013 | 89 (12.0) | 108 (13.5) | ||
| 2014 | 93 (12.6) | 94 (11.7) | ||
| 2015 | 88 (11.9) | 90 (11.2) | ||
| 2016 | 98 (13.3) | 78 (9.7) | ||
| 2017 | 87 (11.8) | 87 (10.9) | ||
|
| ||||
| Top 10, n (%) | No | 635 (85.9) | 765 (95.5) | <0.001 |
| Yes | 104 (14.1) | 36 (4.5) | ||
|
| ||||
| US Residency, n (%) | No | 94 (12.7) | 463 (57.8) | <0.001 |
| Yes | 645 (87.3) | 338 (42.2) | ||
|
| ||||
| US Medical Graduate, n (%) | No | 127 (17.4) | 464 (60.9) | <0.001 |
| Yes | 602 (82.6) | 298 (39.1) | ||
|
| ||||
| Canada Medical Graduate, n (%) | No | 602 (95.3) | 298 (92.0) | 0.058 |
| Yes | 30 (4.7) | 26 (8.0) | ||
|
| ||||
| International Medical Graduate, n (%) | No | 632 (86.7) | 324 (42.8) | <0.001 |
| Yes | 97 (13.3) | 433 (57.2) | ||
|
| ||||
| Visa Status, n (%) | US | 605 (81.9) | 298 (37.2) | <0.001 |
| Canadian | 30 (4.1) | 26 (3.2) | ||
| Other | 104 (14.1) | 477 (59.6) | ||
|
| ||||
| Allopathic Medical Degree, n (%) | No | 15 (2.0) | 27 (3.4) | 0.145 |
| Yes | 724 (98.0) | 774 (96.6) | ||
|
| ||||
| Number of Applications Distributed, median [IQR] | 28 [16, 44] | 9 [4, 19] | <0.001 | |
|
| ||||
| Number of Interviews Marked Completed, median [IQR] | 10 [7, 14] | 1 [0, 5] | <0.001 | |
|
| ||||
| Number of Programs on Applicant’s Rank List, median [IQR] | 10 [6, 14] | 1 [0, 4] | <0.001 | |
|
| ||||
| USMLE Step1 Three Digit Score, median [IQR] | 238 [226, 248] | 227 [214, 239] | <0.001 | |
|
| ||||
| USMLE Step2 Three Digit Score, median [IQR] | 240 [226, 253] | 230 [214, 243] | <0.001 | |
|
| ||||
| USMLE Step3 Three Digit Score, median [IQR] | 225 [214, 235] | 217 [206, 228] | <0.001 | |
Abbreviations: N = total number of applicants who matched in retina, n = number of applicants, IQR = interquartile range (25th - 75th percentile), US = United States, USMLE = United Stated Medical Licensing Exam
p-values generated using Pearson’s chi-squared test for categorical variables and Wilcoxon rank-sum test for continuous variables
Total number (%) of missing observations: 52 (6.6%)
Matched applicants distributed a greater median number of applications (26 vs. 9, p<0.001), completed more interviews (median 10 vs. 1, p<0.001), and ranked more programs (median 10 vs. 1, p<0.001) than unmatched applicants. The benefit of applying to more programs leveled off around 65 applications (Supplemental Figure 1). Above 81 applications, there were large variations with anywhere from 0% to 100% of applicants matching. Two-hundred and thirty-one applicants applied to only 1 program. Of these, 160 (69.3%) completed an interview, 158 ranked a program, and 118 (51.1%) matched. Overall, 4.99% (n=118/2365) of matched applicants only applied to/interviewed at a single program. In addition, 51 of 377 applicants who ranked only 1 program completed more than 1 interview. Of these, 18 (35.3%) did not match. All applicants (n=49, 2.1% of matched applicants) who completed at least 19 interviews matched (Supplemental Figure 2). Compared with unmatched applicants, matched applicants scored higher on USMLE step 1 (median score 238 vs. 227, p<0.001), step 2 (median score 240 vs. 230, p<0.001), and step 3 (median score 225 vs. 217, p<0.001).
Among applicants who matched into retina fellowship during the 2010 to 2017 match cycles, there were yearly variations in the number of interviews completed (range 9–12, p=0.026, Table 3), number of programs ranked (range 8–12, p=0.004), and USMLE step 1 (range 232–241, p=0.004), 2 (range 235–245, p<0.001), and 3 (range 219–231, p<0.001) scores. There were no significant annual variations in the proportion of matched applicants who graduated from a top 10 residency program, U.S. residency program, allopathic medical school, or U.S., Canadian, or international medical school; who held a U.S. visa; or in the number of applications distributed.
Table 3.
Characteristics of Matched Retina Applicants, 2010–2017
| Match Year | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | 2016 | 2017 | p-valuea | |
|---|---|---|---|---|---|---|---|---|---|---|
| Number Matched in Retina+ | 90 | 98 | 96 | 89 | 93 | 88 | 98 | 87 | ||
| Top 10, n (%) | No | 81 (90.0) | 81 (82.7) | 83 (86.5) | 72 (80.9) | 83 (89.2) | 74 (84.1) | 85 (86.7) | 76 (87.4) | 0.626 |
| Yes | 9 (10.0) | 17 (17.3) | 13 (13.5) | 17 (19.1) | 10 (10.8) | 14 (15.9) | 13 (13.3) | 11 (12.6) | ||
|
| ||||||||||
| US Residency, n (%) | No | 10 (11.1) | 14 (14.3) | 15 (15.6) | 13 (14.6) | 11 (11.8) | 9 (10.2) | 14 (14.3) | 8(9.2) | 0.867 |
| Yes | 80 (88.9) | 84 (85.7) | 81 (84.4) | 76 (85.4) | 82 (88.2) | 79 (89.8) | 84 (85.7) | 79 (90.8) | ||
|
| ||||||||||
| US Medical Graduate, n (%) | No | 18 (20.2) | 18 (18.4) | 14 (14.6) | 15 (16.9) | 14 (16.7) | 16 (18.2) | 16 (16.3) | 16 (18.4) | 0.988 |
| Yes | 71 (79.8) | 80 (81.6) | 82 (85.4) | 74 (83.1) | 70 (83.3) | 72 (81.8) | 82 (83.7) | 71 (81.6) | ||
|
| ||||||||||
| Canada Medical Graduate, n (%) | No | 71 (93.4) | 80 (93.0) | 82 (97.6) | 74 (94.9) | 70 (93.3) | 72 (96.0) | 82 (97.6) | 71 (95.9) | 0.733 |
| Yes | 5 (6.6) | 6 (7.0) | 2 (2.4) | 4(5.1) | 5 (6.7) | 3 (4.0) | 2 (2.4) | 3 (4.1) | ||
|
| ||||||||||
| International Medical Graduate, n (%) | No | 76 (85.4) | 86 (87.8) | 84 (87.5) | 78 (87.6) | 75 (89.3) | 75 (85.2) | 84 (85.7) | 74 (85.1) | 0.989 |
| Yes | 13 (14.6) | 12 (12.2) | 12 (12.5) | 11 (12.4) | 9 (10.7) | 13 (14.8) | 14 (14.3) | 13 (14.9) | ||
|
| ||||||||||
| Visa Status, n (%) | US | 72 (80.0) | 80 (81.6) | 82 (85.4) | 74 (83.1) | 72 (77.4) | 72 (81.8) | 82 (83.7) | 71 (81.6) | 0.970 |
| Canadian | 5 (5.6) | 6(6.1) | 2(2.1) | 4 (4.5) | 5 (5.4) | 3 (3.4) | 2 (2.0) | 3 (3.4) | ||
| Other | 13 (14.4) | 12 (12.2) | 12 (12.5) | 11 (12.4) | 16 (17.2) | 13 (14.8) | 14 (14.3) | 13 (14.9) | ||
|
| ||||||||||
| Allopathic Medical Degree, n (%) | No | 1 (1.1) | 4(4.1) | 2 (2.1) | 3 (3.4) | 2 (2.2) | 0 (0.0) | 1 (1.0) | 2 (2.3) | 0.584 |
| Yes | 89 (98.9) | 94 (95.9) | 94 (97.9) | 86 (96.6) | 91 (97.8) | 88 (100.0) | 97 (99.0) | 85 (97.7) | ||
|
| ||||||||||
| Number of Applications Distributed, median [IQR] | 29 [17, 43] | 28 [14, 40] | 27 [14, 45] | 29 [17, 48] | 22 [14, 38] | 28 [16, 41] | 33 [20, 52] | 28 [17, 45] | 0.190 | |
|
| ||||||||||
| Number of Interviews Marked Completed, median [IQR] | 9 [6, 13] | 9 [6, 14] | 11 [8, 14] | 11 [7, 14] | 9 [5, 13] | 11 [7, 14] | 11 [8, 14] | 12 [8, 15] | 0.026 | |
|
| ||||||||||
| Number of Programs on Applicant’s Rank List, median [IQR] | 8 [5, 12] | 9 [5, 13] | 10 [7, 13] | 11 [6, 14] | 8 [4, 12] | 11 [7, 14] | 11 [8, 14] | 12 [8, 15] | 0.004 | |
|
| ||||||||||
| USMLE Stepl Three Digit Score, median [IQR] | 235 [222, 245] | 232 [216, 245] | 236 [225, 249] | 239 [227, 251] | 240 [230, 246] | 239 [227, 250] | 240 [231, 251] | 241 [230, 252] | 0.004 | |
|
| ||||||||||
| USMLE Step2 Three Digit Score, median [IQR] | 236 [223, 248] | 235 [219, 248] | 236 [222, 251] | 237 [219, 249] | 245 [227, 257] | 245 [235, 257] | 244 [231, 256] | 241 [236, 253] | <0.001 | |
|
| ||||||||||
| USMLE Step3 Three Digit Score, median [IQR] | 227 [218, 235] | 219 [209, 228] | 223 [210, 235] | 220 [214, 231] | 227 [213, 237] | 227 [216, 237] | 231 [219, 239] | 228 [218, 235] | <0.001 | |
Abbreviations: n = number of applicants, IQR = interquartile range (25th – 75th percentile), US = United States, USMLE = United States Medical Licensing Exam
p-values generated using Pearson’s chi-squared test for categorical variables and Kruskal-Wallis test for continuous variables
Number (%) of missing observations by year: 2010: 0 (0%), 2011: 0 (0%), 2012: 0 (0%), 2013: 11 (11%), 2014: 13 (12.3%), 2015: 15 (14.6%), 2016: 10 (9.3%), 14 (13.9%)
After controlling for potential covariates, factors associated with matching into retina fellowship were attending a U.S. residency program (odds ratio (OR) 2.08, 95% confidence interval (CI) [1.48, 2.92], p<0.001, Table 4), attending a top 10 ophthalmology residency program (OR 1.74, CI [1.07, 2.84], p=0.026), having an allopathic medical degree (OR 2.39, CI [1.08, 5.33], p=0.032, completing more interviews (OR 1.28, CI [1.23, 1.33], p<0.001), and scoring higher on USMLE step 3 (OR 1.02 [1.01, 1.03], p<0.001). The number of applications distributed and USMLE step 1 and 2 scores were not associated with matching.
Table 4.
Multivariable Logistic Regression Describing Factors Associated with Matching into Retina Fellowship, 2010–2017
| Applicant Characteristics | Odds Ratio | 95% Confidence Interval | p-valuea |
|---|---|---|---|
| US Residency* | 2.08 | [1.48, 2.92] | <0.001 |
| Top 10 Ophthalmology Program* | 1.74 | [1.07, 2.84] | 0.026 |
| Allopathic Medical Degree* | 2.39 | [1.08 5.33] | 0.032 |
| Number of Applications Distributed* | 1.01 | [1.00, 1.02] | 0.134 |
| Number of Interviews Marked Completed* | 1.28 | [1.23, 1.33] | <0.001 |
| USMLE Step1 Three Digit Score | 1.00 | [0.99, 1.01] | 0.623 |
| USMLE Step2 Three Digit Score | 1.00 | [0.99, 1.01] | 0.921 |
| USMLE Step3 Three Digit Score | 1.02 | [1.01, 1.03] | <0.001 |
Abbreviations: US = United States, USMLE = United States Medical Licensing Exam
Reference is “No”
p-values generated using multivariable logistic regression. All covariates used are shown in the table above.
Discussion
Most ophthalmology residents now pursue fellowship training.1 Compared with other subspecialties, retina fellowships have consistently offered the most positions and matched the greatest number of fellowship applicants.3 In the present study, the number of retina fellowship positions offered increased from 136 in 2014 to 160 in 2019, but there was not a concordant increase in the number of positions filled, with up to 37 positions left vacant in 2019. The present data does not elucidate why more positions were unfilled or the breakdown on if these positions were medical or surgical fellowships.
Similar to previous studies describing successful ophthalmology fellowship applicant characteristics,1,8 the present study found that matched retina fellowship applicants shared several characteristics. From 2010 through 2017, almost all (>95%) matched retina applicants had allopathic medical degrees, and more than 75% had graduated from U.S. medical school and residency programs and held a U.S. visa. Interestingly, osteopathic medical graduates comprise about 20% of ophthalmology residents,9 but only 2.7% of applicants in this study. Various factors, such as self-selection or perceived lower odds of matching may play into this discrepancy. Indeed, applicants with an allopathic medical degree were 208% more likely than non-MD’s to match into retina fellowship. However, unmatched applicants were as likely as matched applicants to have an MD, suggesting that graduating from an allopathic medical school may be important, but insufficient, for securing a retina fellowship position.
As was described for all ophthalmology fellowships and glaucoma fellowship,1,8 graduating from a U.S. residency program and completing more interviews were associated with increased odds of matching into retina fellowship. Beyond accepting all interview offers, completing more interviews is largely out of the applicants’ control. We were unable to assess what proportion of interview offers applicants accepted using the available data. Recently, ophthalmology residency programs instituted an interview cap, which was 18 in 2020–21 and 16 in 2021–2022.10 Given that completing 19 interviews for fellowship is costly for the applicant and time-consuming for both the applicant and program, institution of interview capping for fellowships should be considered.
Interestingly, 7.3% of applicants applied to only 1 program, and 1.6% ranked only 1 program despite completing more than 1 interview. Only 49.1% of applicants who restricted themselves to a specific program matched, whereas 76.6% of applicants who applied to, interviewed at, and ranked greater than 1 program matched. These findings suggest that geographical constraints or personal factors may be more critical for some applicants.
For retina applicants, attending a U.S. residency program was relatively less important than for glaucoma applicants (OR 2.08 for retina vs. 9.91 for glaucoma).8 In contrast to these other studies, graduating from a top 10 ophthalmology residency, having an allopathic medical degree, and scoring higher on USMLE step 3 were associated with greater odds of matching in retina. These findings are consistent with a previous survey-based study that found that retina fellowship program directors place more value on the quality of an applicant’s residency program and OKAP scores than glaucoma or cornea program directors.7 At face value, the odds ratio corresponding to USMLE step 3 scores (OR 1.02) is unimpressive. However, scaling up to 10-point intervals reveals that scoring 10 points higher on step 3 is associated with about 6.2 times greater odds of matching. Interestingly, although ophthalmology residents out-perform the national average on step 1 and step 2,11 they score closer to average on step 3.12 It is difficult to determine from this data if high step 3 scores factor into a fellowship selection committee’s decision or if it is merely correlative with the applicant’s academic proficiency. In addition, with Step 1 transitioning to pass/fail in January 2022, admissions committees may place less emphasis on USMLE test scores in the future. Over the last decade, as the percentage of residency programs using Step target scores to filter applicants has declined. According to the annual National Resident Matching Program (NRMP) directory surveys, 60% and 45% of residency programs required target USMLE Step 1 and Step 2 CK scores, respectively, for US MD-candidate applicants to be considered for an interview in 2010.13 In 2021, these values were lower at 45% for Step 1 and 27% for Step 2 CK.14 These data are for residency programs and do not include ophthalmology. Similar data from current ophthalmology fellowship program directors would be interesting to elucidate the true value of USMLE Step 3 scores for ophthalmology fellowship applicants.
These findings are limited by the inclusion of data only from the SF Match. Further, our inability to determine the specialties to which applicants applied or differentiate between medical and surgical retina is a significant limitation. Based on available data we also were not able to directly compare matched retina applicants with unmatched retina applicants. Rather, we compared matched retina applicants to all unmatched applicants who applied to any subspecialty. In addition, given the retrospective nature of the study, we are unable to infer causality for any association discussed. In summary, the number of programs participating in the retina fellowship match rose from 2014 to 2019, but the number of positions filled remained relatively stagnant. Factors associated with matching into retina during the 2010 to 2017 match cycles included graduating from a U.S. residency program, graduating from a top 10 residency program, having an MD, completing more interviews, and scoring higher on USMLE Step 3. All applicants who completed more than 19 interviews matched, but these applicants comprised only 2.1% of matched applicants. Matched applicants completed a median of 10 interviews. Further studies that differentiate between medical and surgical retina will be of interest to prospective applicants in these fields.
Supplementary Material
Acknowledgements/Disclosures
The authors would like to acknowledge Erik Rosales and the Data Resource Committee at AUPO for helping to acquire the SF Match Fellowship Match data. We would also like to thank the Wilmer Biostatistical Department (Core Grant P30-EY01765) for assistance with data analysis.
Financial Support:
This work was supported by the National Institutes of Health (P30EY001765, Wilmer Biostats Core).
Financial Disclosures:
No financial disclosures exist for any author.
Abbreviations:
- SF
San Francisco
- USMLE
United States Medical Licensing Exam
- AUPO
Association of University Professors of Ophthalmology
- OR
odds ratio
Footnotes
Conflict of Interest: No conflicting relationship exists for any author
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Zafar S, Bressler NM, Golnik KC, et al. Fellowship Match Outcomes in the U.S. From 2010 to 2017: Analysis of San Francisco Match. Am J Ophthalmol. 2020. doi: 10.1016/j.ajo.2020.06.008 [DOI] [PubMed]
- 2.Gedde, Budenz DL, Haft P, Lee Y, Quigley HA. Factors affecting the decision to pursue glaucoma fellowship training. J Glaucoma. 2007. doi: 10.1097/01.ijg.0000243474.36213.08 [DOI] [PubMed]
- 3.Dotan G, Karr DJ, Levin AV. Pediatric ophthalmology and strabismus fellowship Match outcomes, 2000–2015. In: Journal of AAPOS.; 2017. doi: 10.1016/j.jaapos.2017.01.004 [DOI] [PubMed]
- 4.Venincasa MJ, Hubschman S, Kuriyan AE, et al. Perceptions of Vitreoretinal Surgical Fellowship Training in the United States. Ophthalmol Retin. 2019. doi: 10.1016/j.oret.2019.05.006 [DOI] [PubMed]
- 5.Chen X, Zafar S, Srikumaran D, et al. Factors Influencing Postgraduate Career Decisions of Ophthalmology Residents. J Acad Ophthalmol. 2020;12(02). doi: 10.1055/s-0040-1715808 [DOI] [Google Scholar]
- 6.Gedde SJ, Budenz DL, Haft P, Tielsch JM, Lee Y, Quigley HA. Factors Influencing Career Choices among Graduating Ophthalmology Residents. Ophthalmology. 2005;112(7). doi: 10.1016/j.ophtha.2005.01.038 [DOI] [PubMed] [Google Scholar]
- 7.Kempton JE, Shields MB, Afshari NA, Dou W, Adelman RA. Fellow Selection Criteria. Ophthalmology. 2009. doi: 10.1016/j.ophtha.2008.12.047 [DOI] [PubMed]
- 8.Burton E, Arsiwala LT, Johnson TV, Srikumaran D, Zafar S, Woreta FA. Applicant Characteristics Associated with Glaucoma Fellowship Match from 2010 to 2017. Ophthalmol Glaucoma. August 2021. doi: 10.1016/j.ogla.2021.08.004 [DOI] [PubMed]
- 9.Loh, Joseph, Keenan JD, Lietman, Naseri A. Predictors of matching in an ophthalmology residency program. Ophthalmology. 2013. doi: 10.1016/j.ophtha.2012.09.028 [DOI] [PubMed]
- 10.Quillen DA, Siatkowski RM, Feldon S. COVID-19 and the Ophthalmology Match. Ophthalmology. 2021;128(2):181–184. doi: 10.1016/j.ophtha.2020.07.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Siatkowski R, Mian S, Cullican S, et al. Probability of Success in the Ophthalmology Residency Match: Three-Year Outcomes Analysis of San Francisco Matching Program Data. J Acad Ophthalmol. 2018. doi: 10.1055/s-0038-1673675 [DOI] [PMC free article] [PubMed]
- 12.USMLE Score Interpretation Guidelines. https://www.usmle.org/pdfs/transcripts/USMLE_Step_Examination_Score_Interpretation_Guidelines.pdf.
- 13.National Resident Matching Program. Results of the 2010 NRMP Program Director Survey.; 2010. https://www.nrmp.org/wp-content/uploads/2021/11/2021-PD-SurveyReport-for-WWW.pdf. Accessed May 15, 2022.
- 14.National Resident Matching Program. Results of the 2021 NRMP Program Director Survey.; 2021. https://www.nrmp.org/wp-content/uploads/2021/11/2021-PD-Survey-Report-for-WWW.pdf. Accessed May 15, 2022.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.