If you ask busy residents what they do with their free time, they are likely to reply, “What free time?” Historically, graduate medical education research has focused solely on what residents do during their working or on-duty hours. More recently, concerns over patient safety have directed investigators' attention to the effects of sleep deprivation and fatigue on clinical performance, personal health, and medical errors.1–3 To date, however, researchers have paid little attention to what residents do with their nonwork and nonsleep time.
The 2003 Accreditation Council for Graduate Medical Education (ACGME) 80-hour-per-week limits for resident work hours were, in part, intended to give residents more opportunity for sleep in an effort to counter rising concerns that medical errors are frequently the result of sleep deprivation and fatigue.4 Our own research in 1999 demonstrated that when residents moved from their first postgraduate year (PGY) to their second, they gained an extra hour of nonwork time per day. On average, however, the PGY-2 residents reported using only 13.8 minutes of this time for extra sleep; they used the rest for other activities.1,5
The results of our 1999 national survey provided data from residents in multiple specialties on their work and sleep hours in the context of the full 168 hours available weekly.5 Large differences were evident in the average self-reported weekly work hours across the 22 specialties in our survey, ranging from a high of 110.6 hours per week for neurosurgery to a low of 56.7 hours for pathology. When we added the self-reported average weekly sleep hours to these work hours for each specialty, we found equally large differences in the time remaining (i.e., residents' free time)—about which we had no information. We found wide variation in the hours of time remaining, ranging from as few as 23.8 hours per week of nonwork and nonsleep time for neurosurgical residents to 63.2 hours per week for pathology residents—The difference between the two is nearly threefold.5 PGY-1 residents averaged 44.7 hours a week of such residual time, whereas PGY-2 residents reported 49.9 hours of such time, a significant difference of 5.16 hours per week (P < .001).1 Since the advent of the 2003 ACGME rules limiting work hours in a given week, the amount of nonwork, nonsleep time available per week has increased substantially.6,7
We are not aware of other research that looks specifically at how residents spend this nonwork, nonsleep time. Some of the literature discussing the residency training experience has used the term “lifestyle.” The authors using this term, however, have focused on residents' preferences for specialties such as dermatology, ophthalmology, and radiology. Many perceive that the specialties, combining the benefits of shorter training requirements, higher compensation, reduced work hours, and predictable work demands, offer a more “controllable lifestyle.”8–11 Others have framed these lifestyle concerns as part of residents' search for improved “work–life balance.”12 All of this research focuses more on how residents' expectations affect their careers and specialty choices than on what residents actually do with their unstructured time.
The most relevant data on this “other life” of residents are reported by Gander and colleagues,13 who, like ourselves, employed a large, multispecialty survey of “junior physicians” in New Zealand. They addressed the desire for work–life balance in these physicians-in-training by asking about the effects of “work-related fatigue risk” on their lives outside of work. They report the numerous consequences of fatigue risk on problems in social life, home life, personal relationships, and other commitments. Respondents' written comments indicated difficulties with social isolation, self-care, personal relationships, and caring for children, as well as with studying.
Other studies have attempted to address recent changes in the residency experience by examining specific on-duty experiences such as workload, operative opportunities, work-related stress, educational experiences, and medical errors.14,15 For example, Mourad and colleagues16 have reported that some residents defer indirect patient care duties, such as chart review and dictation, to after hours, in violation of current work hours limitations. Brasel and colleagues17 include “off-duty hours,” and Jamal and colleagues18 include “extracurricular activities” in their studies of work-related variables. Neither set of investigators, however, clearly identifies these variables or specifies where and when they occur. Their overall concern is more about how changes in work hours affect working-time activities, rather than what residents do outside of work.
Still other researchers have studied resident well-being in an effort to address specific issues related to residents' personal health and work-related stress.12,19,20 These studies have expressed concern about the high prevalence of depression and anxiety among residents, but they have not fully explored the range of adaptive or coping responses used by residents or what residents may do outside of work hours to mitigate the effects of stressful work circumstances.21,22
Taken together, these studies may help the academic medicine community to understand some of the changing conditions of residency training, but they shed little light on what residents do during their nonwork, nonsleep time. Largely absent from the literature is any accounting of residents' specific activities outside of the workplace. We expect that some of these activities involve attending to the tasks of daily life (e.g., clothing, food, shelter, bills) as well as to the demands of ongoing relationships with friends and family. Of equal concern, we do not know enough about how residents may use their off-duty time for self-care, personal enjoyment, rest, relaxation, and mental and emotional repair. How residents spend this time could measurably influence their work, learning, satisfaction with residency, professional behavior, and sleep patterns.
Noting this absence in the literature as well as in our own research, in the fall of 2009, we undertook a multiinstitutional, multispecialty resident survey, carrying forward prior lines of inquiry, and adding several new questions that examine residents' experience with and attitudes toward the 2003 ACGME duty hours limits.4 With this new survey, we hoped to fill two major gaps in the academic medicine community's understanding of the residency experience: (1) In what activities do residents participate during their nonwork and nonsleep hours? and (2) How much does the individual resident's experience vary within and between programs? This report focuses on the first of these questions.
Because the immediate purpose of the survey was to collect data on variations in individual residency experiences both within and between programs, we needed very high response rates and established a predefined criterion of at least 70%. In addition, we restricted our sample to residents in four major specialties: internal medicine, obstetrics–gynecology, pediatrics, and surgery. We selected these because they are among the largest specialties.
We contacted 27 hospitals representing 48 programs from across the country. We purposely selected the hospitals for their diversity in size, governance (i.e., public versus private status), and geographic location.
We surveyed residents about their recently completed first and second year of training. The survey instrument was quite detailed, containing 50 questions and providing 189 self-reported data points. It covered a variety of issues: the number of hours residents work and sleep, residents' activities during the remaining time, information about various elements in residents' learning and training experiences (e.g., their sources of learning, their view of the quality of supervision, the frequency with which they have committed five different types of medical errors), their personal health, psychological abuse, and behavioral changes. In addition, we included three widely used, validated scales assessing sleepiness (the Epworth Sleepiness Scale),23 depression (the Center for Epidemiological Studies–Depression–10 scale),24,25 and anxiety (the Spielberger Anxiety Scale).26
To examine trends over time, we included some items on the questionnaire that were identical to items in our previous large, random, national, multispecialty resident surveys in 1989 and 1999.1,5 In addition, we added several new questions deemed germane to current issues in graduate medical education, including the index question for this report: “During your past year of residency, outside of working hours, about how often did you….” This was followed by a list of 10 possible activities and a four-point rating scale (1 = “less than once a week”; 4 = “almost daily”). Listed activities were seeing a movie, watching television, reading for leisure, reading professional journals, moonlighting, using the Internet, spending time with friends, spending time with family, doing household tasks, and exercising.
The authors contacted the designated institutional officials and program directors of the selected specialty training programs at the identified institutions and asked for their cooperation in distributing the surveys. Through both the cover letter and telephone conversations, we described the purpose of the survey, explained assurances of confidentiality and anonymity, and specified the required response rate. To encourage participation, we offered program directors the opportunity to receive their own aggregate data if they so wished. We reminded nonresponding program directors and coordinators by both letter and phone. We secured external, expedited institutional review board (IRB) approval from the American Institutes of Research. Respondents were able to answer the survey online using SurveyMonkey.com (Palo Alto, California) or by completing a mail-in paper-and-pencil questionnaire.
To allow some institutions to consult their own IRBs or to accommodate competing survey demands, we conducted data collection from late August through October 2009. Because the primary intent of the survey was to explore variations in individual resident experiences rather than to collect averages, we explicitly required a targeted sample of programs with high response rates.
We conducted analyses using SPSS release version 18.0.0 (SPSS, Inc., Chicago, Illinois, 2009), and we conducted basic descriptive statistical analyses, K-cluster analyses, factor analyses, one-way analyses of variance, and stepwise discriminant function analyses.
Of the 27 institutions we initially invited, 5 declined to participate or failed to respond. Six other institutions failed to meet our stringent response rate requirements, leaving a final sample of 16 institutions comprising 36 programs, 13 of which had response rates of 100%. Per U.S. Census Bureau definition of region, 3 institutions (8 programs) were located in the Mid-Atlantic region, 5 institutions (10 programs) were in the Midwest, 4 institutions (7 programs) were in the West, and 4 institutions (11 programs) were in the South.
The final sample consisted of 759 residents, of whom 634 (83.5%) responded. Respondents included 216 residents from 9 internal medicine programs, 109 residents from 9 surgery programs, 80 residents from 8 obstetrics–gynecology programs, and 229 residents from 10 pediatric programs. Just over 60% (378 of 617) of the sample were female, 51.1% (288 of 619) were married, 20.4% (126 of 618) lived in a household with children, and 23.3% (144 of 617) were either U.S. citizen or non-U.S.-citizen international medical graduates (IMGs).
PGY-1 internal medicine residents averaged 50.2 hours (standard deviation [SD] = 8.6) hours of nonwork, nonsleep time per week; pediatric residents averaged 49.0 (SD = 9.1) nonwork, nonsleep hours per week; obstetrics–gynecology residents averaged 47.8 (SD = 7.1) hours per week; and surgery residents averaged 42.6 (SD = 11.0) hours per week. The figures increased slightly for PGY-2 residents: Internal medicine residents averaged 51.8 (SD = 10.0) nonwork, nonsleep hours per week; pediatric residents averaged 50.0 (SD = 10.1) nonwork, nonsleep hours per week; obstetrics–gynecology residents averaged 47.7 (SD = 7.4) hours per week; and surgery residents averaged 45.1 (SD = 7.0) hours per week.
For all residents, Internet use, occurring almost daily, was the most commonly reported activity, followed in frequency by watching television (Figure 1). By contrast, spending time moonlighting or watching a movie were uncommon activities; almost all residents reported that these occurred less than once a week (Table 1). Both reading for leisure and reading professional journals were also fairly infrequent, whereas doing household chores, spending time with friends and/or family, and getting physical exercise averaged about once a week. These averages are misleading, however. Some residents reported spending several times or more per week with friends (n = 152/620; 24.5%) or with family (n = 218/620; 35.2%), doing household tasks (n = 198/622; 31.8%), or getting physical exercise (n = 195/621; 31.4%).
As a means of assessing relationships among the different reported activities, we entered responses to the 10 activity items listed in Table 1 into a series of K-cluster analyses. We then examined solutions for two, three, four, five, and six clusters. We selected the three-cluster solution because it met both the criteria of having relatively similar sample sizes for each cluster and of offering the most substantively interpretable solution. Table 2 lists the average reported activities for the three clusters. Using these average scores, we named the three clusters by focusing on the activities with the highest scores within each cluster, relative to the other two clusters. We named Cluster 1, whose members reported spending the most time with friends, high Internet use, the most time exercising, and the most time watching television, the “Friend-Focused” cluster. We deemed Cluster 2, whose members also reported high scores for Internet use and watching television, along with the highest comparative scores for time spent with family and doing household tasks, the “Family-Focused” cluster. Finally, the comparatively lower scores reported for every activity on the list led us to call Cluster 3 the “Low Activity” cluster.
Low Activity residents
Tables 3 and 4 show the relationship between these three activity clusters and other variables assessing the residency training experience. For almost every variable, the Low Activity cluster is noticeably different from the other two. Low Activity residents reported significantly less satisfaction with the residency experience overall, as well as with what they learned. They also reported statistically significantly higher levels of depression, anxiety, and stress (P < .001). Two other findings supporting this sense of general dysphoria are that Low Activity cluster residents were also more likely than the other respondents to feel that time for essential clinical tasks was inadequate and to think that issues related to their residency experience were problematic.
For Low Activity residents, weekly work hours were longer and weekly sleep hours fewer. They also reported statistically higher Epworth sleepiness scores and saw themselves as more sleep deprived (P < .001). These same residents were more likely to report both that they had been belittled or humiliated and that they had fewer friends in their residency program.
Although Low Activity residents did not report fewer weekly hours with attending physicians (not shown), they reported that both attending physicians and supervisory residents were less important as sources of learning than did the other two clusters. Reports of having made a significant workload-related medical error were higher among Low Activity residents (P = .054), as were reports of increased use of caffeine (P = .01). However, these residents were not more likely to take medications to go to sleep, experience significant weight changes, or report increased use of alcohol compared with the residents in the other two clusters.
Demographically, Low Activity residents were, on average, significantly older (P < .001) than their peers and significantly more likely to be IMGs (P < .001). Half of them were married (n = 106), and 26% (n = 50) had children. In comparison, about 80% of Family-Focused and 25% of Friend-Focused residents were married, and about a quarter of Family-Focused and just a few Friend-Focused residents had children at home. We found significant differences across specialties, with 40.8% (n = 42/103) of surgical residents, 39.5% (n = 30/76) of obstetrics–gynecology residents, 34.3% (n = 71/207) of internal medicine residents, and 24.2% (n = 53/219) of pediatric residents being classified as belonging to the Low Activity cluster (χ2 = 17.95, df = 6, P = .006).
Residents from the Low Activity cluster were found in all residency programs, ranging from 71% (5 of 7 residents) in the program with the greatest percentage to 11% (1 of 9 residents) in the program with the lowest percentage.
Stepwise discriminant function analysis
Finally, to estimate which of the examined factors might be most important in differentiating among the three clusters, we entered all of the variables in the survey into a stepwise discriminant function analysis. This analysis produced a model composed of two functions based on whether or not residents were married, what their anxiety scores were, whether they had children in their households, how many total weekly hours they worked, whether they had graduated from an international medical school, and how many friends they reported in their residency program. Using these two functions, we were able to classify 58.2% (352 of 605) of the residents into the correct cluster, although the classification was better for the Friend-Focused cluster (67.4%, 155 of 230) and the Family-Focused cluster (69.3%, 124 of 179) than for the Low Activity cluster (37.2%, 73 of 196).
Residents engage in a variety of different activities outside of their work and sleep hours. Which activities they engage in, and how much time per week they spend in each activity, varies across individual residents.
Our findings suggest a typology for understanding residents' activities outside of work using two dimensions: friend focused versus family focused and high activity versus low activity. Using these dimensions, we identified three separate clusters of residents: those whose activities appeared to be primarily friend focused, those whose activities seemed family focused, and those who showed low activity levels. Low Activity residents reported lower satisfaction with their residency experience, higher levels of depression and anxiety, longer work hours, more sleepiness, more experiences of being belittled or humiliated, and fewer friends in their residency program. A model based on six factors could classify residents into one of these three defined clusters with some level of reliability.
The most common activities reported by residents—spending time on the Internet and watching television—were solitary activities that residents can do at flexible times and for varying durations of time. These activities lend themselves to the irregular schedules of residents and can easily fit into the spaces between more rigidly scheduled activities. The negligible reports of moonlighting may reflect that the residents responding to our survey were in high-intensity specialties and that moonlighting is uncommon for PGY-1 and PGY-2 residents.5,27
Most residents reported spending some time socializing each week with friends, family, or both, and this time seems to be beneficial. The Family-Focused cluster in particular reported the lowest scores for depression, anxiety, and stress. Nonwork social contact certainly gives residents a personal life beyond the work environment and may serve as an outlet to reduce stress. The Low Activity residents, however, reported not only less social contact but also less activity of all types. The issue here is not simply one of being asocial but, seemingly, of being disconnected.
Although the Low Activity residents reported longer work hours and less sleep, it may be tempting to conclude that they were simply more engaged and, therefore, chose to work rather than to take advantage of the potential free time. The fact that these residents exhibited more dysphoric reactions on all other indicators, however, seems to refute such a conclusion. Our findings are somewhat reminiscent of those of Tanz and Charrow,28 who studied a group of pediatric residents claiming to have larger, heavier workloads and more “bad luck” than their colleagues. These authors found no evidence of any real workload difference and concluded that these “black cloud” residents worked inefficiently and created extra work for themselves. This cluster may also include many of the so-called “strugglers” or “problem residents”—individuals with ongoing difficulties—identified by program directors.29,30
The Low Activity cluster was not homogenous and included a wide variety of people. Half were married, some had children, both genders were represented, and some were IMGs. What distinguished these residents was not who they were but what they did—or, rather, what they failed to do. Simply put, they reported substantially less time in both individual and social activities of all types, demonstrating a pervasive lack of engagement in all activities outside of structured residency work time. Although these Low Activity residents were more common in certain specialties (i.e., surgery, obstetrics–gynecology), they were found, to a greater or lesser degree, in every one of the programs in our sample. This suggests that it is not the characteristics of the program but, rather, the actions of the residents themselves that determine their cluster designation.
We do not see low activity, by itself, as an intrinsic problem. Rather, we see it as a risk factor for impaired learning and emotional health. Certainly, not all Low Activity residents were depressed or unsatisfied; nor were all depressed people in the Low Activity cluster. Still, the finding that, in the aggregate, Low Activity residents evinced more dysphoric responses suggests that the lack of outside activities may represent another of the “broken windows” indicators we have written about elsewhere.31 The notion of broken windows is drawn from the observation that a neighborhood is in trouble when property is neglected.32,33 Broken windows themselves are not the real problem; rather, they are a concrete sign of underlying social disorganization which is less easily identified. Low activity is not a problem in and of itself, but it may be a red flag indicating that some residents may not be satisfactorily coping or learning. At the very least, we suspect that low activity outside of work suggests that the residents are not developing life patterns characterized by personal and professional balance.34 Unlike the Family-Focused and Friend-Focused residents, who spend potentially supportive time with family or friends, Low Activity residents seem to be disengaged and, by all indications, not doing as well.
Ours was not a random sample of residents. The underlying purpose of our survey was to study individual variation within and between residency programs. To do this, we needed to secure very high levels of participation from residents in four of the largest specialties. Thus, the reports of these residents do not necessarily reflect the selections of residents from less demanding specialties who have more available free time. On the other hand, they probably give us a better picture of what residents do when their condensed free time constrains their choices.
As we have done for our previous survey reports,1,5,7,27,31 we acknowledge the caveat that our findings are all based on self-report data. The high level of participation, however, suggests that the residents were interested, willing partners in the study.
These results are based on residents' reports of their usual weekly activities. The work life of a resident, however, is likely to be highly variable from week to week. As residents move from service to service, they encounter differing organizational cultures, personnel, workloads, structures, expectations, and demands. Because the practice in many programs is to try to alternate more stressful rotations with less demanding ones, the notion of a “usual” week must be taken as a subjective average.
We also must point out that we asked residents how often they engaged in the listed activities, not the actual amount of time they spent. Although some activities reported as infrequent could possibly have occurred for a longer duration (e.g., taking a full day with family might be reported as only once a week), we read the scaled answers as subjective reports of how much or how little the residents engaged in these activities.
Finally, residents were offered a preselected list of activities, which limited their choices. Although the list of activities was derived from consultation with residents both in survey design and in pilot testing, respondents were not given the opportunity to suggest alternative activities while completing the survey. Other activities, such as studying, commuting, conducting a romance, and accessing personal health care, should be included in future surveys.
Implications and suggestions for future work
Future work should assess the actual amounts of time residents allocate to each activity, in view of the large number of hours devoted to studying, as reported by Gander and colleagues.13 In addition, because time allocations shift from week to week, a detailed assessment of these shifts and the factors that influence them would add to the academic medicine community's understanding of how residents structure their lives.
Until the recent implementation of work hours rules, the question of what residents did when they were not working or sleeping was probably a moot point. Residents in many specialties had little or no time for much beyond the demands of work and the necessities of sleep. But, as work hours have been reduced, residents currently have more free or spare time than in any time in recent medical education history.6,7 Our data suggest that what they do, or fail to do, with that time may have real implications for graduate medical education as well as for residents' health and well-being.
Although some may believe that how residents use their nonwork, nonsleep time is largely under their personal direction and control, we think this is not the case. In particular, we wonder how much free time residents actually have, given external and work obligations. Like all human beings, they have pressing needs for food, shelter, and economic transactions that require their attention. This supposed free time is further encumbered by the responsibilities of marriage, children, other family members, and friends, whose needs and demands are seldom trivial. How well residents balance the competing demands for their time has a direct impact on how well they are able to perform at work and attend to their learning. Further, alongside these competing demands, residents may require additional time to “repair” the physical and emotional strain of the training experience.13
We speculate that part of the motivation for reducing work hours stems from the changing demographic profile and psychosocial expectations of today's residents.13 As mentioned in the discussion about lifestyle issues, many residents express an increased desire for greater personal work–life balance.35 Rather than work hours reduction creating the potential for increased activities, we suspect that the desire for time to pursue personal activities has become a key dynamic driving the reduction in work hours.
Training the physicians of tomorrow requires looking beyond the workday and even beyond the implications of sleep deprivation to consider what residents are doing and what they could be doing with their recently acquired unstructured time.7 If the best physicians are to be not merely technically competent but also capable of maintaining well-balanced, ethical lives,36 we should give more consideration to helping them achieve this balance. All work and no play may not make the best physician, but limiting work hours apparently does not guarantee that physicians will play. We hope this report will begin to change the conceptual framework37 from a tunnel-vision focus on work and sleep hours to a more complete consideration of the life experience of residents.
The authors wish to thank all of the participating residents and program directors for their interest and involvement.
1. Baldwin DC Jr, Daugherty SR. Sleep deprivation and fatigue in residency training: Results of a national survey of first and second year residents. Sleep. 2004;27:217–223.
2. Landrigan CP, Rothschild JM, Cronin JW, et al.. Effect of reducing interns' work hours on serious medical errors in intensive care units. N Engl J Med. 2004;351:1838–1848.
3. Ulmer C, Wolman DM, Johns MME, eds. Resident Duty Hours: Enhancing Sleep, Supervision, and Safety. Washington, DC: National Academies Press; 2009.
5. Baldwin DC Jr, Daugherty SR, Tsai R, Scotti MJ Jr. A national survey of residents' self-reported work hours: Thinking beyond specialty. Acad Med. 2003;78:1154–1163.
6. Staiger DO, Auerbach DA, Buerhaus PI. Trends in the work hours of physicians in the United States. JAMA. 2010;303:747–753.
7. Baldwin DC Jr, Daugherty SR, Ryan P, Yaghmour NA. Changes in resident work and sleep hours 1999 to 2009: Results from a survey of 4 specialties. J Grad Med Educ. 2010;2:656–658.
8. Schwartz RW, Haley JV, Williams C, et al.. The controllable lifestyle factor in students' attitudes about specialty selection. Acad Med. 1990;65:207–210.
9. Dorsey ER, Jarjoura D, Rutecki GW. Influence of controllable lifestyle on recent trends in specialty choice by US medical residents. JAMA. 2003;290:1173–1178.
10. Hudkins JR, Helmer SD, Smith RS. General surgery resident practice plans: A workforce for the future? Am J Surg. 2009;198:798–803.
11. DeZee KJ, Maurer D, Colt R, et al.. Effect of financial remuneration on specialty choice of fourth-year US medical students. Acad Med. 2011;86:187–193.
12. Ratanawongsa N, Wright SM, Carrese JA. Well-being in residency: A time for temporary imbalance? Med Educ. 2007;41:273–280.
13. Gander P, Briar C, Garden A, Purnell H, Woodward A. A gender-based analysis of work patterns, fatigue, and work/life balance among physicians in postgraduate training. Acad Med. 2010;85:1526–1536.
14. Papp KK, Stoller EP, Sage P, et al.. The effects of sleep loss and fatigue on resident-physicians: A multi-institutional, mixed-method study. Acad Med. 2004;79:394–406.
15. Fletcher KE, Underwood W III, Davis SQ, Mangrulkar SR, McMahon LF Jr, Saint S. Effects of work hour reduction on residents' lives: A systematic review. JAMA. 2005;294:1088–1100.
16. Mourad M, Vidyarthi AR, Hollander H, Ranji SR. Shifting indirect patient care duties to after hours in the era of work hours restrictions. Acad Med. 2011;86:586–590.
17. Brasel KJ, Pierre AL, Weigelt JA. Resident work hours: What they are really doing. Arch Surg. 2004;139:490–494.
18. Jamal MH, Rousseau MC, Hanna WC, Doi SAR, Meterissian S, Snell L. Effect of the ACGME duty hours restrictions on surgical residents and faculty: A systematic review. Acad Med. 2011;86:34–42.
19. Ratanawongsa N, Wright SM, Carrese JA. Well-being in residency: Effects on relationships with patients, interactions with colleagues, performance, and motivation. Patient Educ Couns. 2008;72:194–200.
20. Cohen JS, Patten S. Well-being in residency training: A survey examining resident physician satisfaction both within and outside of residency training and mental health in Alberta. BMC Med Educ. 2005;5:21. http://www.biomedcentral.com/1472-6920/5/21
. Accessed December 30, 2011.
21. Prins JT, van der Heijden FM, Hoekstra-Weebers JE, et al.. Burnout, engagement and resident physicians' self-reported errors. Psychol Health Med. 2009;14:654–666.
22. Peterlini M, Tibério IF, Saadeh A, Pereira JC, Martins MA. Anxiety and depression in the first year of medical residency training. Med Educ. 2002;36:66–72.
23. Johns MW. A new method for measuring daytime sleepiness: The Epworth Sleepiness Scale. Sleep. 1991;14:540–545.
24. Radloff LS. The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1:385–401.
25. Andresen EM, Malmgren JA, Carter WB, Patrick DL. Screening for depression in well older adults: Evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). Am J Prev Med. 1994;10:77–84.
26. Spielberger CD, Gorsuch RL, Lushene RE, Vagg PR, Jacobs AG. Manual for the State-Trait Anxiety Inventory. Palo Alto, Calif: Consulting Psychologists Press; 1983.
27. Baldwin DC Jr, Daugherty SR. Moonlighting and indebtedness reported by PGY2 residents: It's not just about money! Acad Med. 2002;77(10 suppl):S36–S38.
28. Tanz RR, Charrow J. Black clouds: Work load, sleep, and resident reputation. Am J Dis Child. 1993;147:579–584.
29. Paice E. Identification of the underperforming surgical trainee. ANZ J Surg. 2009;79:180–184.
30. Durning SJ, Cleary T, Sandars J, Hemmer P, Kokotailo P, Artino AR. Perspective: Viewing “strugglers” through a different lens: How a self-regulated learning perspective can help medical educators with assessment and remediation. Acad Med. 2011;86:488–495.
31. Baldwin DC Jr, Daugherty SR. Interprofessional conflict and medical errors: Results of a national multispecialty survey of hospital residents in the US. J Interprof Care. 2008;22:573–586.
32. Wilson JQ, Kelling GL. Broken Windows: The Police and Neighborhood Safety. New York, NY: Atlantic Monthly; 1982.
33. Kelling GL, Coles CM. Fixing Broken Windows: Restoring Order and Reducing Crime in Our Communities. New York, NY: Martin Kessler; 1996.
34. Novack DH, Suchman AL, Clark W, Epstein RM, Najberg E, Kaplan C. Calibrating the physician: Personal awareness and effective patient care. JAMA. 1997;278:502–509.
35. Defoe DM, Power ML, Holzman GB, Carpentieri A, Schulkin J. Long hours and little sleep: Work schedules of residents in obstetrics and gynecology. Obstet Gynecol. 2001;97:1015–1018.
36. Lopez L, Katz JT. Perspective: Creating an ethical workplace: Reverberations of resident work hours reform. Acad Med. 2009;84:315–319.
37. Schwartz A, Pappas C, Bashook PG, et al.. Conceptual frameworks in the study of duty hours changes in graduate medical education: A review. Acad Med. 2011;86:18–29.
The American Institutes for Research granted approval for this study following an expedited review.
The results of a similar cluster analysis solution of preliminary data were presented at the March 2010 Accreditation Council for Graduate Medical Education annual conference.© 2012 Association of American Medical Colleges