Job burnout, defined as a psychological syndrome characterized by emotional exhaustion, depersonalization, and feelings of decreased personal accomplishment,1 is common among internal medicine (IM) residents2 and may lead to poor academic performance,3,4 depression,5 suicidality,6 self-perceived medical errors,7 needle-stick injuries, and motor vehicle accidents.8 Potential risk factors for the development of burnout in IM residents include personality type,2 limited performance feedback,2 educational loan debt,3 and duty hours.9,10
Not surprisingly, excessive duty hours contribute to fatigue in IM residents, and decreased work hours may improve measures of sleepiness.7,11 In addition, burnout seems to correlate with sleep deprivation in medical students12 and IM residents,13 whereas, conversely, recovery from burnout may improve sleep continuity.14 Given that emotional exhaustion is a component of job burnout, researchers have hypothesized that fatigue resulting from excessive duty hours contributes to burnout.9,10 In 2003, the Accreditation Council for Graduate Medical Education (ACGME) adopted duty hours standards for all specialties. Studies evaluating the impact of these limitations found that emotional exhaustion levels modestly decreased while burnout levels overall were unaffected.9,10 Five years later, the Institute of Medicine (IOM) produced a report, “Resident Duty Hours: Enhancing Sleep, Safety and Supervision,”15 which recommended a protected sleep period of 5 hours during any work shift longer than 16 hours to reduce the risk of fatigue-related errors committed by residents in the hospital during prolonged duty periods.16 The IOM report acknowledged the paucity of data on optimizing duty hours for physicians in training, but argued that the evidence on the hazards of fatigue-related performance errors in other professions likely extended to medicine. In 2011, the ACGME modified the duty hours standards to limit continuous duty of first-year residents to 16 hours.17
The level of burnout amongst first-year IM residents from multiple academic institutions was previously reported by our group prior to the 2011 modification.2 The more limiting 2011 duty hours restrictions (DHRs) created an opportunity to measure the impact of duty hours on the development of job burnout and to examine important mediators such as inadequate sleep. We hypothesized that duty hours limitations would lead to decreased job burnout in first-year IM residents.
We invited all preliminary and categorical first-year residents beginning an IM or combined medicine–pediatrics internship in 2011 from the following training programs to participate: Icahn School of Medicine at Mount Sinai (New York, New York), Perelman School of Medicine at the University of Pennsylvania (Philadelphia, Pennsylvania), and Massachusetts General Hospital (Boston, Massachusetts). Eligible residents were not offered incentives to encourage participation. Two additional centers from our original 2008–2009 cohort did not participate in 2011–2012. For purposes of comparison, we excluded data from those two centers from our analysis.
Our survey instrument has been described previously.2 Surveys included demographic and resident characteristic questions, and validated tools to measure burnout and sleep deprivation. Pre and post surveys varied slightly in order to capture information unobtainable prior to the start of training (e.g., average number of patients admitted per call).
Burnout was measured using the Maslach Burnout Inventory (MBI), a validated 22-item instrument that measures emotional exhaustion, depersonalization, and sense of personal accomplishment.1 For each of the three burnout domains, a subscore was generated. Though the MBI is commonly used to measure burnout in resident physicians, some variability exists in interpreting MBI score responses.13,18 For purposes of comparison with the greatest number of studies in the literature, we defined burnout using the most common definition, a high subscore on either the emotional exhaustion or depersonalization subscales.
The validated Epworth Sleepiness Scale (ESS) was used to measure fatigue. Respondents rate on a 4-point scale the likeliness of their falling asleep in eight different scenarios. A summative score is calculated, and excessive sleepiness is defined by a total of greater than 10 points.19
Each participating program adapted to the revised standards and service demands differently. These programmatic modifications are highlighted in Table 1. All programs adhered to mandated shift length and time-off requirements.
We administered surveys by printed questionnaire similarly in the 2008–2009 and 2011–2012 cohorts.2 All participants were informed of the risks of completing the survey and had the opportunity to decline participation. For both cohorts, we administered the first survey in June, just prior to the start of training during intern orientation. Second surveys were administered near the end of the first year of training, between April and June of 2009 and 2012, with slight variation based on institution. We invited residents to participate in the study at program meetings, before or after clinic sessions, or through intrainstitutional mailing.
Using unblinded administrative coordinators from each institution, individual residents’ pre and post responses were paired while maintaining anonymity. The coordinators had no access to survey responses, and we had no access to the unique identifiers assigned to each participant. Collaborating institutions removed all personal information from completed surveys prior to transfer for data entry and analysis. Data analysis for the 2011–2012 cohort was completed by January 2013. Each participating institution named above obtained institutional review board approval prior to survey administration.
We reviewed respondents’ demographic data using univariate distributions to compare pre and post profiles from two separate classes of first-year residents, the initial group surveyed in 2008–2009 and the more recent 2011–2012 cohort. Multiple-comparison chi-square tests were conducted to identify statistically significant differences in proportions (P < .05) of burnout across survey items and year of administration. The data were dichotomized with the desired responses (agree or disagree) versus all other responses combined. We used Fisher exact test when cell sizes were small (n < 5). We used SAS for Windows statistical software, version 9.1 (SAS Institute, Inc., Cary, North Carolina) for data analysis. All data were analyzed with the intention to identify correlations between survey items, burnout incidence, and late-internship burnout prevalence. This dataset will likely be used for additional future studies attempting to correlate job burnout with other potential outcomes, such as measures of professionalism and medical errors.
In the 2011–2012 cohort, 188 first-year residents from the three participating institutions were eligible to participate in our study; 181 (96%) completed the initial survey, 133 (71%) completed the second survey, and 128 (68%) completed both surveys. Of 180 eligible residents in the 2008–2009 cohort, 111 (62%) completed both surveys. Some respondents chose not to answer every item on the questionnaire.
Demographic characteristics of the 2008–2009 and 2011–2012 cohorts are shown in Table 2. The groups were similar in terms of most measured characteristics. However, members of the 2011–2012 cohort had fewer females (51/121 [42%] versus 60/108 [56%], P =.04), were somewhat less likely to have taken time off (> 1 year) before the start of medical school (31/118 [26%] versus 44/107 [41%], P = .02), and included slightly more foreign medical graduates (6/120 [5%] versus (0/106 [0%], P = .01). More residents self-reported a history of anxiety in the 2011–2012 cohort; however, this finding did not reach statistical significance (9/120 [8%] versus 2/106 [2%], P = .07). In both the 2008–2009 and 2011–2012 cohorts, categorical residents were slightly though significantly more likely to complete both surveys compared with other residents. We found no other significant demographic differences between the residents who completed both surveys compared with those who completed just the first.
Prevalence and incidence
The prevalence and incidence of burnout in the two cohorts is reported in Table 3. Year-end prevalence (92/123 [75%] versus 91/108 [84%], P = .08) and new (incident) cases (59/87 [68%] versus 55/68 [81%], P = .07) of burnout in the 2011–2012 cohort were no different than in the previous cohort, even after the implementation of the revised DHRs. We performed site-level analyses to explore the possibility of varying results between participating programs and found that there were no significant differences in burnout prevalence or incidence across the three institutions in either cohort.
Sleep and patient service characteristics
Associations between self-reported sleepiness, service characteristics, and job burnout are reported in Table 4. There was no significant difference in end-of-the-year excessive ESS sleepiness scores between the 2011–2012 and 2008–2009 cohorts (72/122 [59%] versus 71/108 [66%], P = .29). More residents in the 2011–2012 cohort than the 2008–2009 cohort (2011–2012 versus 2008–2009) (52/123 [42%] versus 10/69 [14%], P < .01) and more of the residents who developed new burnout over the course of the year in the 2011–2012 than the 2008–2009 cohort (2011–2012 versus 2008–2009) (29/59 [49%] versus 5/34 [15%], P < .01) reported caring for more than 8 patients on inpatient services.
Though there was no difference between the 2008–2009 and 2011–2012 cohorts in terms of self-reported hours worked, residents within the 2011–2012 cohort who developed burnout were more likely than those who remained burnout-free to report working > 70 hours per week (49/59 [83%] versus 18/28 [64%], P = .05).
Job burnout among first-year IM and combined medicine/pediatrics resident physicians from the three university-based residency programs we studied did not significantly change after the implementation of the 2011 revised DHRs. The only common significant change since 2008 across all four participating institutions was the 16-hour DHR implemented in 2011 (see Table 1).
Despite DHR implementation, excessive sleepiness levels (our measure of fatigue) were found to be similar in the two cohorts. This finding may be explained by the similar number of hours worked before and after the mandate. Though residents in the 2011–2012 cohort worked fewer consecutive hours than those in the 2008–2009 cohort, they reported a similar number of total hours worked per week and a similar number of days off per month. It is likely that sleepiness may not accurately reflect fatigue, with fatigue and sleepiness being different constructs.20 Excessive daytime sleepiness, as measured in our study by the ESS,19 primarily assesses the likelihood that an individual will fall asleep in a given situation. The change in extended first-year IM resident duty hours from 30 to 16 has likely changed the work intensity for trainees of all levels through shifting of tasks from interns to residents and work compression for first-year residents who now have less time to perform similar amounts of clinical work. Such work compression leads to a change in work intensity that might contribute to fatigue without affecting ESS scores. Interestingly, first-year residents in our study reported caring for more patients after the implementation of shift length restrictions, supporting the theory that work intensity may be greater.
Though there were lower rates of burnout in IM residents after the implementation of the 2011 ACGME DHRs, the difference did not reach significance, and the majority of first-year residents still developed burnout by year’s end. The persistently high rates of resident burnout likely reflect the multifactorial nature of job stress in doctors-in-training, which includes factors such as work intensity, lack of control over schedules, participation in life-and-death decision making, witnessing tragedy, and the cumulative effect of long duty hours.9,21 Participating in the care of patients with unpredictable courses and the impact of the hidden curriculum may lead to emotional experiences that challenge the values of young trainees and add to the stresses of training.22 There may also be other unanticipated negative consequences of duty hours standards that continue to drive burnout in residents. For example, the majority of surgery residents in a cohort surveyed after the 2011 duty hours standards reported decreased coordination of care and no change in levels of fatigue, while a third of them experienced at least weekly symptoms of depersonalization and emotional exhaustion.23
Some argue that job burnout may be an acceptable consequence of training if it promotes quality of care, and that some aspects of the burnout syndrome, such as desensitization to traumatic patient experiences, could be formative to professional development.24 However, it seems more likely that the continued high rate of burnout is potentially harmful to the trainee’s mental state5,6 and to patient care.5,7,25,26 Further study should attempt to elucidate the components and correlates of burnout that place the physician-in-training at greatest risk both for personal complications of burnout and for medical errors. Despite the possible modest decline in burnout observed with the more restrictive 2011 DHRs, the optimal number and organization of duty hours is not clear. Dutch IM residents, for example, who work fewer hours than their U.S. counterparts, continue to experience burnout, though at lower rates than in the United States.27
This study has several important limitations. The use of a historical control group raises the possibility that factors other than the revised duty hours standards influenced rates of burnout. Although all three sites reduced duty hours in accordance with the ACGME mandate, each site employed different changes to the structure of their training program (see Table 1). Because there were similar rates of burnout development after the 2011 DHRs were implemented across the sites, it is reasonable to conclude that work limitations themselves do not significantly lower rates of burnout. The inclusion of multiple sites strengthens the ability to draw this conclusion. Nonetheless, the analysis performed identified correlations only, making it difficult to draw conclusions regarding causality. Furthermore, given the poor response for a few of the variables reported in Table 4, there could potentially be a nonresponse bias for some of these specific outcomes. Also, the response rate in the 2008–2009 cohort was lower than in the 2011–2012 cohort. The impact of the lower response rate on our results is not clear. Burnt-out residents may be less motivated to respond, which would lead us to underestimate the rate of burnout, or they may be more inclined to respond with the hopes that their participation will lead to positive change, which would lead us to overestimate burnout rates. In addition, controversy persists regarding the optimal definition for burnout13,18; the definition most commonly used in studies of resident burnout was applied in our cohort to make our findings most comparable with those from many other studies. Additionally, some attributes associated with burnout inventories (e.g., exhaustion, subjugation of personal attachment) may be related to positive aspects of physician development in ways that are not measurable in this study design. Finally, the measures of fatigue20 and job demands21 that were used are largely self-reported and could be enhanced to better assess the potential mediators of the correlation between burnout and duty hours.
The revised ACGME DHRs of 2011 created an opportunity to measure the impact of DHR on burnout. Despite a number of unanswered questions, it is clear that doctors-in-training undergo a personal and professional transformation during residency that poses significant challenges and hazards. The balance between resident well-being, substantive education,4,9 and patient safety is yet to be established. This multi-institutional study found that the implementation of 2011 duty hours standards was not associated with significantly lower rates of burnout among first-year IM residents, and that unacceptably high rates of burnout persist. Future research should focus on additional correlates of burnout in order to optimize the training environment to maximize patient safety and to enhance the professional development of tomorrow’s physician workforce.
1. Maslach C, Jackson SE, Leiter MP Maslach Burnout Inventory Manual. 19963rd ed. Palo Alto, Calif Consulting Psychologists Press
2. Ripp J, Babyatsky M, Fallar R, et al. The incidence and predictors of job burnout in first-year internal medicine residents: A five-institution study. Acad Med. 2011;86:1304–1310
3. West CP, Shanafelt TD, Kolars JC. Quality of life, burnout, educational debt, and medical knowledge among internal medicine residents. JAMA. 2011;306:952–960
4. West CP, Shanafelt TD, Cook DA. Lack of association between resident doctors’ well-being and medical knowledge. Med Educ. 2010;44:1224–1231
5. Shanafelt TD, Bradley KA, Wipf JE, Back AL. Burnout and self-reported patient care in an internal medicine residency program. Ann Intern Med. 2002;136:358–367
6. van der Heijden F, Dillingh G, Bakker A, Prins J. Suicidal thoughts among medical residents with burnout. Arch Suicide Res. 2008;12:344–346
7. West CP, Tan AD, Habermann TM, Sloan JA, Shanafelt TD. Association of resident fatigue and distress with perceived medical errors. JAMA. 2009;302:1294–1300
8. West CP, Tan AD, Shanafelt TD. Association of resident fatigue and distress with occupational blood and body fluid exposures and motor vehicle incidents. Mayo Clin Proc. 2012;87:1138–1144
9. Gopal R, Glasheen JJ, Miyoshi TJ, Prochazka AV. Burnout and internal medicine resident work-hour restrictions. Arch Intern Med. 2005;165:2595–2600
10. Goitein L, Shanafelt TD, Wipf JE, Slatore CG, Back AL. The effects of work-hour limitations on resident well-being, patient care, and education in an internal medicine residency program. Arch Intern Med. 2005;165:2601–2606
11. 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
12. Mazurkiewicz R, Korenstein D, Fallar R, Ripp J. The prevalence and correlations of medical student burnout in the pre-clinical years: A cross-sectional study. Psychol Health Med. 2012;17:188–195
13. Rosen IM, Gimotty PA, Shea JA, Bellini LM. Evolution of sleep quantity, sleep deprivation, mood disturbances, empathy, and burnout among interns. Acad Med. 2006;81:82–85
14. Ekstedt M, Söderström M, Akerstedt T. Sleep physiology in recovery from burnout. Biol Psychol. 2009;82:267–273
15. Ulmer C, Wolman DM, Johns MME Resident Duty Hours: Enhancing Sleep, Supervision, Safety. 2008 Washington, DC Institute of Medicine
16. 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
18. Dyrbye LN, West CP, Shanafelt TD. Defining burnout as a dichotomous variable. J Gen Intern Med. 2009;24:440
19. Johns MW. A new method for measuring daytime sleepiness: The Epworth sleepiness scale. Sleep. 1991;14:540–545
20. Shen J, Barbera J, Shapiro CM. Distinguishing sleepiness and fatigue: Focus on definition and measurement. Sleep Med Rev. 2006;10:63–76
21. Demerouti E, Bakker AB, Nachreiner F, Schaufeli WB. The job demands–resources model of burnout. J Appl Psychol. 2001;86:499–512
22. Billings ME, Lazarus ME, Wenrich M, Curtis JR, Engelberg RA. The effect of the hidden curriculum on resident burnout and cynicism. J Grad Med Educ. 2011;3:503–510
23. Antiel RM, Reed DA, Van Arendonk KJ, et al. Effects of duty hour restrictions on core competencies, education, quality of life, and burnout among general surgery interns. JAMA Surg. 2013;148:448–455
24. Nash IS. Resident burnout. Ann Intern Med. 2002;137:698–700
25. Fahrenkopf AM, Sectish TC, Barger LK, et al. Rates of medication errors among depressed and burnt out residents: Prospective cohort study. BMJ. 2008;336:488–491
26. West CP, Huschka MM, Novotny PJ, et al. Association of perceived medical errors with resident distress and empathy: A prospective longitudinal study. JAMA. 2006;296:1071–1078
27. Prins JT, Hoekstra-Weebers JE, Gazendam-Donofrio SM, et al. Burnout and engagement among resident doctors in the Netherlands: A national study. Med Educ. 2010;44:236–247