Data are expressed as a percentage of same-sex classmates within each year, and groups reaching significance are indicated by P value.
eTable 1. Survey Sent to Residents in Our Multi-institutional Study
eTable 2. Survey Sent to Program Directors in Our Multi-institutional Study
eTable 3. Reasons for Considering Leaving Residency
eTable 4. Reasons for Deciding to Stay in General Surgery
Gifford E, Galante J, Kaji AH, Nguyen V, Nelson MT, Sidwell RA, Hartranft T, Jarman B, Melcher M, Reeves M, Reid C, Jacobsen GR, Thompson J, Are C, Smith B, Arnell T, Hines OJ, de Virgilio C. Factors Associated With General Surgery Residents’ Desire to Leave Residency ProgramsA Multi-institutional Study. JAMA Surg. 2014;149(9):948-953. doi:10.1001/jamasurg.2014.935
General surgical residency continues to experience attrition. To date, work hour amendments have not changed the annual rate of attrition.
To determine how often categorical general surgery residents seriously consider leaving residency.
Design, Setting, and Participants
At 13 residency programs, an anonymous survey of 371 categorical general surgery residents and 10-year attrition rates for each program. Responses from those who seriously considered leaving surgical residency were compared with those who did not.
Main Outcomes and Measures
Factors associated with the desire to leave residency.
The survey response rate was 77.6%. Overall, 58.0% seriously considered leaving training. The most frequent reasons for wanting to leave were sleep deprivation on a specific rotation (50.0%), an undesirable future lifestyle (47.0%), and excessive work hours on a specific rotation (41.4%). Factors most often cited that kept residents from leaving were support from family or significant others (65.0%), support from other residents (63.5%), and perception of being better rested (58.9%). On univariate analysis, older age, female sex, postgraduate year, training in a university program, the presence of a faculty mentor, and lack of Alpha Omega Alpha status were associated with serious thoughts of leaving surgical residency. On multivariate analysis, only female sex was significantly associated with serious thoughts of leaving residency (odds ratio, 1.2; 95% CI, 1.1-1.3; P = .003). Eighty-six respondents were from historically high-attrition programs, and 202 respondents were from historically low-attrition programs (27.8% vs 8.4% 10-year attrition rate, P = .04). Residents from high-attrition programs were more likely to seriously consider leaving residency (odds ratio, 1.8; 95% CI, 1.0-3.0; P = .03).
Conclusions and Relevance
A majority of categorical general surgery residents seriously consider leaving residency. Female residents are more likely to consider leaving. Thoughts of leaving seem to be associated with work conditions on specific rotations rather than with overall work hours and are more prevalent among programs with historically high attrition rates.
Despite structural changes to residency programs during the past decade, including adoption of the 80-hour and then 16-hour rules, resident attrition continues to be a problem facing general surgery programs across the country.1 Modern attrition rates for general surgery residents remain between 3% and 5.1% annually and total 19% during the course of a 5-year to 7-year residency program.2,3 Most of this attrition is voluntary and not secondary to poor resident performance.4 Despite investigations that have examined the cohort of residents that leaves surgical training, little attention has been focused on residents who consider quitting residency but ultimately elect to stay. Although the annual Accreditation Council for Graduate Medical Education resident survey evaluates work hours and attempts to gauge resident experience, it does little to address risk factors for quitting.3
We conducted a multi-institutional survey of categorical general surgery residents to identify how often they seriously consider leaving residency. The objective was to identify what factors are associated with an increased desire to leave residency and which factors mitigate that desire. We also sought to determine whether the desire to quit was more prevalent among programs with historically higher attrition rates. Ultimately, our aim was to better understand the significance of various stressors in the life of a resident so that programs can adopt changes that go beyond broad work hour rules to improve resident satisfaction and retention.
The study was approved by the Human Subjects Committee of the Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center. The study was deemed exempt from participant informed consent because the survey was anonymous and voluntary. A 52-question survey was created by the program director (C.d.V.) at Harbor-UCLA Medical Center (eTable 1 in the Supplement) in conjunction with other program directors (including J.G.) in the study and with select surgical residents (E.G. and C.R.). The anonymous survey consisted of demographic information, followed by a question about whether the resident had ever seriously considered leaving the general surgery residency and, if so, how often in each postgraduate (PG) year he or she had completed. Residents who indicated that they had never wanted to quit completed the survey at this point. Those residents who at some point had seriously considered quitting went on to respond to questions using a Likert-type scale (score range, 1-5). They were asked to rate which elements of their program or personal life influenced their thoughts of quitting. Thereafter, they were asked what factors were most helpful in mitigating those thoughts and moving forward with their training. Concurrently, a second survey was sent to the program directors at each of the participating institutions (eTable 2 in the Supplement). The program director was asked to indicate how many categorical general surgery residents were available for survey participation, as well as the attrition rate for the preceding 10 years and whether attrition for each resident was voluntary or involuntary. As a follow-up step, program directors were contacted individually and were asked the sex-specific attrition and the sex breakdown of the graduating residents from 2004 to 2013.
The anonymous survey was sent to 371 categorical general surgery residents at 13 residency programs. Based on prior communication that indicated high interest in study participation, programs were chosen to represent different regions of the United States (West, Southwest, Midwest, and Northeast) and various training models (university program, independent program, or hybrid program [university affiliation without an on-site university or medical school]). Our survey was distributed at the conclusion of the 2013 academic year to capture those residents who had participated in at least 1 year of training. The questionnaire was administered using online survey software (SurveyMonkey; https://www.surveymonkey.com/). To link the historical attrition rate of each program with the responses from that program, a unique survey link was created for each program. Responses were collected in an anonymous fashion directly from the residents from June 10, 2013, to August 18, 2013. To maintain anonymity, program directors were not given access to the results of their individual program.
Statistical analysis of resident and program responses was exported into a native SAS format using a computer program (DBMS/Copy; Dataflux Corporation). Analyses were conducted using statistical software (SAS, version 9.3; SAS Institute Inc). The analysis of the overall study population and the survey results is descriptive and simply uses proportions and medians. For the main objective of the study, the cohort of residents who had never thought of quitting was compared with those who had ever thought of quitting. In addition, a comparison was made between residents who were at programs with a historically high attrition rate vs residents who were at programs with a historically low attrition rate, as defined by a cutoff value of 19%. The numerical determination of high-attrition vs low-attrition programs was defined based on a survey by Yeo et al5 that demonstrated a 5-year to 7-year cumulative attrition rate of 19% for a general surgery residency program. For univariate analyses, continuous numerical variables were compared using the nonparametric Wilcoxon rank sum test and were reported as medians (interquartile ranges [IQRs]). Categorical or nominal variables were compared using χ2 test or Fisher exact test, as appropriate. P < .05 was considered statistically significant. Factors that were found to be significantly predictive of a desire to quit on univariate analysis were subjected to a multivariate analysis. Because of the concern that residents from a single program may be more likely to respond similarly to one another, we used the PROC GENMOD (general linear modeling) procedure. Finally, a stratified analysis of sex differences in survey response was also performed to further explore why sex was an independent predictor of the outcome of having ever thought of quitting residency.
Two hundred eighty-eight residents (77.6%) responded to the survey. Of the respondents, 176 (61.1%) were men, and 112 (38.9%) were women (Table 1). Postgraduate year ranged from 1 to 7 (median, 3; IQR, 2-4), with elongations in training attributable to years spent for research. Most (220 residents [76.4%]) were receiving training at a university program. Only 23.6% were training at a hybrid program (34 residents [11.8%]) or an independent program (34 residents [11.8%]). The median age of respondents was 30 years (IQR, 29-32 years). Two hundred twenty residents (76.4%) had a significant other, while 76 residents (26.4%) had 1 or more children. Thirty residents (10.4%) underwent remediation at some point during their training. Almost all the respondents (256 [88.9%]) identified general surgery as the only specialty pursued in the residency match.
One hundred sixty-seven residents (58.0%) indicated that they had seriously considered leaving surgical residency at some point during their training. The median frequency at which residents considered leaving residency was “a few times in the year.” As a percentage of respondents per training year, residents considered leaving training most often in PG year 1 (45.8%) and year 2 (41.4%) and during research years (35.4%).
With respect to the question, “Have you ever seriously considered leaving residency?” the factors that had the highest association with this finding (as determined by the percentages who “agree” or “strongly agree”) were sleep deprivation on a specific rotation (50.0%), an undesirable future lifestyle (47.0%), and excessive work hours on a specific rotation (41.4%) (eTable 3 in the Supplement). With respect to the question, “I decided to stay in general surgery because…,” the factors that had the highest percentage of affirmative responses were support from family or significant others (65.0%), support from other residents (63.5%), and perception of being better rested (58.9%) (eTable 4 in the Supplement).
On univariate analysis, factors that were significantly associated with the consideration to leave residency were older age, female sex, PG year, training in a university program, lack of Alpha Omega Alpha status, and the presence of a faculty mentor (Table 2). In addition, those who had considered leaving were less likely to respond that they had an accurate perception of general surgery before entering residency (P = .02). On multivariate analysis, only female sex was significantly associated with serious thoughts of leaving residency training (odds ratio, 1.2; 95% CI, 1.1-1.3; P = .003). The consideration to leave training by women was independent of their specific residency program. A separate analysis of the cohort of 30 respondents who underwent remediation did not demonstrate any significant difference from the respondents as a whole.
Comparing the responses of women and men, women were more likely to continue to have serious thoughts of leaving as residency progressed, with women reporting thoughts of leaving more often than men in PG years 2 and 3, research years, and chief years (Figure). When comparing the responses of men and women about factors influencing this consideration, women were more likely to identify sleep deprivation on a specific rotation (P = .005) and difficult interaction with specific faculty (P = .05) as influencing those thoughts. Factors that women associated with the decision to stay in general surgery were perception of being better rested and loss of friends and camaraderie if they left (Table 3).
The 10-year mean attrition rate for the 13 programs was 14.4% (range, 5.0%-29.3%). Three programs had a high (≥19%) median 10-year attrition rate of 27.8% (IQR, 21.6%-29.3%). The remaining 10 programs had a low (<19%) median 10-year attrition rate of 8.4% (IQR, 6.9%-9.9%) (P = .04). Eighty-seven residents left training between 2004 and 2013, and 39 of them were women (median, 46.4%; IQR, 30.8%-65.0%). Adjusting for the ratios of men to women in the 10 years of graduating residents, women left training more often than men, with an odds ratio of 1.9 (95% CI, 1.2-3.0; P = .005). No difference was found in female attrition rates between the high-attrition and low-attrition programs (P = .68). In addition, no difference was observed in the ratios of men to women at high-attrition vs low-attrition programs (P = .25). Comparing responses in our prospective survey, residents from high-attrition programs (86 residents) were more likely to seriously consider leaving residency than those from low-attrition programs (202 respondents) (odds ratio, 1.8; 95% CI, 1.0-3.0; P = .03).
This multi-institutional survey of 288 categorical general surgery residents at 13 residency programs sought to determine how often residents seriously considered leaving residency and to identify what factors were associated with this response. Overall, 58.0% of respondents seriously considered leaving their training, with a median frequency of a few times a year. The survey interviewed a broad variety of residents from large university programs, independent general surgery programs, and hybrid programs, with an overall response rate of 77.6%. The sex distribution of the respondents was similar to overall trends in general surgery residency, with 61.1% male and 38.9% female (compared with 60% and 33% distributions nationally).6 On multivariate analysis, only female sex was associated with serious thoughts of leaving residency. Furthermore, such thoughts persisted within the female cohort throughout residency, whereas male counterparts were less likely to report such sentiments as they advanced through training. The validity of the survey results is strengthened by the additional finding that residents from programs with historically high attrition rates were more likely to seriously consider leaving than residents from programs with historically low attrition rates.
The present survey revealed several notable findings about why residents consider leaving surgical residency. Residents were most likely to cite sleep deprivation on a specific rotation and excessive work hours on a specific rotation, but not work hours overall, as influencing their desire to leave. With respect to sleep deprivation and work hours, this may suggest that it is the workload of an individual rotation that residents find most stressful rather than the rigor of the entirety of general surgery. Therefore, a potential remedy may be to identify these high work-hour rotations and modify them accordingly.
A more difficult issue to address is the finding in the survey that 47.0% of residents who expressed a desire to leave cited an undesirable future lifestyle as a major factor that influenced this sentiment. Practicing surgeons continue to experience high levels of work-home conflicts and burnout.7 In a study7 of more than 7000 surgeons, greater than 50% reported a work-home conflict. Such conflict was cited as a major cause of burnout, depression, and substance abuse. In addition, changes in the training structure may contribute to the negative outlook by residents in our study. In a 2009 American Board of Surgery survey,8 the findings indicated that 27% of respondents were concerned they would be unprepared for practice at the end of residency, and 64% reported that they would have to pursue specialty training following residency. Increased training duration and need for further specialization, along with persistent negative stereotypes of the surgical lifestyle, continue to impair the attitudes of residents. It is doubtful that the residents’ negative impression of a surgeon’s future lifestyle will change unless the rigors of surgical life following residency are altered.
During the past 20 years, the Residency Review Committee has implemented significant requirements for residency programs, many of which would be anticipated to improve attrition rates. The 80-hour workweek has not had such an effect. A comparison of attrition at 215 general surgery training programs from before vs immediately after the institution of the 80-hour workweek demonstrated no change in voluntary resident withdrawal (53% vs 48%).9 Indeed, subsequent studies3,5,10 have debunked the theory that work hour changes will decrease resident attrition. We believe that our survey findings highlight the fact that a desire to leave training may not be affected by job rigor alone but rather program-specific or rotation-specific factors or dissatisfaction with a future career in general surgery.
Previous studies3- 5,11 have identified factors that predict surgery resident attrition. In a single-center survey,11 residents who left cited lifestyle, family, and marital issues, alongside wanting to pursue a specialty other than surgery. Resident age and sex have not been predictive of attrition in some studies.3,5 Predictive of leaving were junior position in training (PG years 1 and 2), dissatisfaction with operative and program experience, lack of camaraderie among residents, and the sentiment that training is too long,3 whereas poor performance during residency has not been predictive.4 In a national study,3 residents who left training in PG years 1 and 2 were almost 3 times more likely to report considering leaving in the prior year. The present study reinforces the notion, as supported by Sullivan et al,3 that thoughts of quitting must be taken seriously and should be a red flag for possible attrition. These findings also suggest that efforts to reduce attrition would be best directed toward junior residents, improving their operative volume and building camaraderie.
In the present study, women were more likely to report wanting to leave residency. That sentiment persisted throughout their training, whereas it did not for their male counterparts. In addition, analysis of the 10-year attrition rates at the 13 programs surveyed demonstrated that a higher percentage of women than men left training. Previous studies1- 3,5,10 demonstrated that women and men view training differently. In a survey of interns,5 women believed that surgery was not a welcoming career for them because of difficulty in maintaining family life, limited flexible training, and lack of role models. In another survey by Viola et al,12 female residents stated that lifestyle was the most important factor in choosing to pursue specialty training. While married male residents and married male residents with children had higher levels of work satisfaction, married women with children reported the highest levels of residency-induced strain on family life. These findings may explain why women in our survey continued to consider leaving residency throughout the duration of training and underscores the importance of supporting female residents through the difficult balance between motherhood and professional life.
Another obstacle facing female trainees is the paucity of female mentors in academic surgery.3,13- 16 Data from the American Board of Surgery In-Training Examination exit survey8 for 2008 found that women were more likely to report that their program was less supportive and that they did not have a faculty member to turn to when difficulties arose. In addition, female applicants entering surgical programs are more aware of sex disparity than male applicants and applicants pursuing nonsurgical fields.15 Striving to increase the number of female faculty members within training programs and refining the mentor-mentee relationship with incoming residents may improve the outlook and productivity of future female surgeons.
Our multi-institutional questionnaire has some limitations. First, we relied on voluntary participation, which is subject to response bias. It is unknown whether systematic differences existed between respondents and nonrespondents. However, our survey response rate of 77.6% was robust. Second, only residents who reported a desire to leave residency filled out the second portion of the survey. We did not record how residents who did not want to leave perceived their residency program. This related to our study objective of capturing the motivations of residents who actually considered quitting but not gauging resident satisfaction as a whole. Third, comparisons between high-attrition and low-attrition programs should be interpreted with caution because only 3 programs represented the high-attrition cohort. Fourth, while the attitudes of our respondents correlate well with actual attrition rates at the interviewed programs, our findings do not represent factors that actually cause attrition but rather may encourage such thought processes to occur and negatively influence the training experience for all involved.
The training of surgical residents is a long and arduous process that necessitates an immense investment of time for the trainee and the faculty. As such, resident attrition is a tremendous loss for all involved parties. In this multi-institutional survey of surgical residents, a majority seriously considered leaving their training, and most had such thoughts more than once. Given that prior investigations indicate that surgical residents who think of quitting are more likely to subsequently do so,3 the survey results herein are sobering. With the increasing number of women entering surgical training, the fact that female sex predicted thoughts of quitting in the present study is similarly concerning. Surgical training programs should take heed of these findings and work in a cooperative fashion to address factors that increase residents’ desire to leave surgical residency.
Accepted for Publication: April 4, 2014.
Corresponding Author: Christian de Virgilio, MD, Department of Surgery, Harbor-UCLA Medical Center, 1000 W Carson St, Campus Box 25, Torrance, CA 90502 (firstname.lastname@example.org).
Published Online: July 30, 2014. doi:10.1001/jamasurg.2014.935.
Author Contributions: Dr de Virgilio had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Gifford, Jarman, Hines, de Virgilio.
Acquisition, analysis, or interpretation of data: Gifford, Galante, Kaji, Nguyen, Nelson, Sidwell, Hartranft, Jarman, Melcher, Reeves, Reid, Jacobsen, Thompson, Are, Smith, Arnell, de Virgilio.
Drafting of the manuscript: Gifford, Nguyen, de Virgilio.
Critical revision of the manuscript for important intellectual content: Gifford, Kaji, Nelson, Sidwell, Hartranft, Jarman, Melcher, Reeves, Reid, Jacobsen, Thompson, Are, Smith, Arnell, Hines, de Virgilio.
Statistical analysis: Gifford, Galante, Kaji, Nguyen.
Administrative, technical, or material support: Gifford, Nelson. Jarman, Reeves, Reid, Smith, Hines, de Virgilio.
Study supervision: Galante, Nelson, Hartranft, Jarman, Are, de Virgilio.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented at the 2014 Annual Meeting of the Pacific Coast Surgical Association; February 17, 2014; Dana Point, California.
Correction: This article was corrected on August 14, 2014, to fix an incorrect word in the Abstract and Results.