Resident work hours have been a subject of intense debate for several decades. Long work hours are associated with resident fatigue, burnout, depression, and patient care errors.1–4 However, shifts to reduce hours have led to more frequent handoffs and compression of the workday.5 In response to concerns about long work hours of physicians-in-training, the Accreditation Council for Graduate Medical Education (ACGME) implemented sequential work hours limitations (WHLs). The first of these, in July 2003, specified that residents could not work more than 80 hours per week, more than 24 hours per shift (+ 6 hours for education and handoffs), or take call more than one night out of every three. For most programs, this represented a profound change in residency training. Prior to this, more than half of interns nationwide reported working more than 80 hours per week.6 In response to recommendations from the Institute of Medicine to further reduce resident work hours and increase direct supervision, the ACGME implemented even more restrictive WHLs in July 2011. These latest WHLs specified that interns were limited to 16-hour shifts, and second-year residents and beyond were limited to 28-hour shifts (see Supplemental Digital Appendix 1 at http://links.lww.com/ACADMED/A441 for an overview of WHLs).
Many hoped that the implementation of progressive WHLs would improve both patient safety and resident well-being. A systematic review of studies in 2010 found an association between shorter shift length and a reduction in medical errors7; however, a study using a large Medicare data set found no difference in mortality or readmissions after the 2011 WHLs were implemented.8 Currently, the large, randomized, multicenter iCOMPARE study is in progress and aims to shed more light on the impact of the latest WHLs on patient safety.9
The impact of progressive WHLs on internal medicine (IM) resident well-being may be more difficult to assess. A meta-analysis of survey studies suggests that the 2003 WHLs may have contributed to reduced burnout incidence among residents.10 However, a multisite study found that the 2011 WHLs had no effect on year-end burnout prevalence.11 A systematic review in 2015 found that depression was still prevalent among residents, and that the incidence may be increasing over time.12 Finally, a large, multisite, multispecialty study on resident perceptions of the 2011 WHLs found that only interns perceived the WHLs to have a positive impact on quality life (61.8%) compared with 13.9% of senior residents reporting a positive effect and 49.7% reporting a negative effect.13
To our knowledge, no previous analysis has measured changes in validated resident well-being metrics at one institution across three time points spanning two iterations of WHLs. We therefore chose to compare scores on resident well-being questionnaires from 2001, 2004, and 2012 at our multisite, university-based IM residency program. We surveyed residents in the spring of 2012 to measure perceived effects of the 2011 duty hours limitations. Previous surveys done at the same time of year in 2004 (following implementation of the 2003 WHLs) and 2001 (prior to WHLs beyond an 80-hour workweek recommendation) at our institution allowed for longitudinal comparison of resident well-being, career satisfaction, and perceived educational quality.2,14 Given prior studies showing a limited impact of reduced work hours on rates of burnout and depression in residency,10–12,14 we hypothesized that resident well-being in 2012 would remain similar to that reported in 2001 and 2004 despite implementation of the 2011 WHLs.
Study design and setting
To evaluate the impact of the 2011 WHLs, we mailed a 58-item, anonymous, self-administered survey to residents’ homes in the spring of 2012. The survey addressed the perceived impact of the latest WHLs, and several validated questionnaires assessing resident well-being. Many sections of the survey were purposefully identical to surveys mailed to residents in the same residency program in spring 2001 and spring 2004. In our program, the vast majority of our inpatient rotations included 24+ hour call for both interns and senior residents prior to the 2011 WHLs.
All residents in the University of Washington Affiliated Hospitals Internal Medicine Residency Program were eligible to participate. Within this program, all residents rotated through three separate hospitals in Seattle, and some residents spent their entire second year in Boise, Idaho. As an incentive to participate, residents who returned a separately mailed postcard indicating that they had completed a survey were eligible for a drawing for one of three $100 gift certificates. The University of Washington Human Subjects Institutional Review Board approved the study. Information regarding this study and about available mental health resources for residents was distributed to the residents before, during, and after the study.
The 2012 survey consisted of validated markers of resident well-being; questions assessing resident perception of the impact of the 2011 WHLs on resident well-being, patient care, and resident education; and questions regarding resident preferences for WHLs structure in the future. Participants were permitted to skip any questions that they did not feel comfortable answering. Details of 2001 and 2004 study procedures have been described previously.2,14 Missing values were handled consistently across the three surveys.
The 22-item Maslach Burnout Inventory15 (MBI) was used to measure burnout. According to convention, burnout was defined as a high score on either the depersonalization or emotional exhaustion subscale. Depression was assessed using the validated two-question PRIME-MD depression screen.16,17 Career satisfaction was assessed using questions that have been used and described previously.2,14 All of the questions assessing resident well-being in 2012 were identical to the questions asked in 2001 and 2004.
The survey also included questions about the perceived impact of the latest WHLs on patient care, resident education, and resident well-being. These questions were identical to questions from the 2004 survey. Participants were asked to respond using five-point Likert scales describing agreement (1 indicates strongly disagree; 2, somewhat disagree; 3, neutral; 4, somewhat agree; and 5, strongly agree) or the nature of the effect (1 indicates strong negative effect; 2, somewhat negative effect; 3, little or no effect; 4, somewhat positive effect; and 5, strong positive effect). For analytic purposes, questionnaire responses were dichotomized, with a response > 3 scored as “positive.”
Finally, there was a question specific to only the 2012 survey asking what work hours structure residents would prefer moving forward.
We used descriptive analyses to compare resident characteristics across the three study periods (2001, 2004, and 2012). We also compared resident characteristics across study periods using chi-square tests. We used extended chi-square tests to test our a priori hypothesis that resident questionnaire responses would not differ across the three time points. In post hoc exploratory analyses, we separately compared 2001 questionnaire responses with 2012 responses and compared 2004 responses with 2012 responses on items with statistically significant overall results using pairwise chi-square tests. Analyses were performed using Stata statistical software, version 12.1 (StataCorp, College Station, Texas). For all tests, a two-sided α of 0.05 was used.
Participant demographic characteristics were similar across the three time points (Table 1). The survey response rate was 66% (112/170) in 2012, compared with 73% (118/162) in 2004 and 76% (115/151) in 2001.
Validated markers of resident well-being
Regarding burnout as assessed by the MBI, overall 61% (68/112) of 2012 residents met burnout criteria compared with 64% (75/118) of surveyed residents in 2004 and 76% (87/115) in 2001, P = .039. In post hoc pairwise comparisons, this difference was statistically significant between 2001 and 2012 (P = .016) but not between 2004 and 2012 (P = .657) (Table 3). In 2012, the mean scores on the MBI subscales were as follows: emotional exhaustion subdomain 24.5 (± 10.2), depersonalization subdomain 10.7 (± 5.8), and personal accomplishment subdomain 37.7 (± 5.7).
Regarding the two-item PRIME-MD depression screening, in 2012, 31% (35/112) of residents screened positive for depression, which was decreased from 55% (65/118) in 2004 and 45% (52/115) in 2001, P = .001. The lower proportion of residents with a positive depression screen was statistically significant when comparing 2001 with 2012 (P = .03) and 2004 with 2012 (P < .0001).
Regarding career satisfaction, 79% (89/112) of 2012 residents responded positively to the question, “At this point in your residency how happy are you with your career choice?” This was significantly higher than residents in 2001 (66% [76/115] responding positively, P = .024) but similar to 2004 (80% [94/118] responding positively, P = .97). (See Tables 2 and 3.)
Resident perceptions of WHLs impact on well-being, education, and patient care
When asked about their perceptions of WHLs, only 23% (26/112) of the 2012 residents overall thought that the 2011 WHLs had a positive effect on their well-being (Figure 1). Senior residents (R2sS/R3s), in particular, did not feel that the latest WHLs improved well-being, with 56% (38/68) reporting a negative impact on well-being, compared with 32% (14/44) of interns (R1s) endorsing a negative impact (Figure 1). This is in contrast to overall results from the 2004 survey, when 84% (99/118) of residents thought that the 2003 WHLs had a positive impact on their well-being.
When asked about the impact of the 2011 WHLs on patient care, the majority of residents felt that the impact was either negative (41% [46/112]) or neutral (42% [47/112]). Again, the R2s/R3s were more likely to report a perceived negative impact on patient care (51% [35/68]) compared with the R1s (25% [11/44]). Overall, these results are similar to the results from the 2004 survey. In 2004, 37% (44/118) reported a negative impact of the 2003 WHLs on patient care, and 34% (40/118) reported a neutral impact.
Regarding the perceived impact of the 2011 WHLs on resident education, most residents (62% overall, or 69/112) reported a negative impact, including the majority of R1s (52% [23/44]) and R2s/R3s (68% [46/68]). The trend in this result is similar to that in the 2004 study. In 2004, 47% (55/118) reported a negative impact of the 2003 WHLs on their own education.
Resident preferences for future WHLs
When asked which work hours scenario they would prefer moving forward, 69% (47/68) of senior residents favored reverting to the pre-July 2011 system of duty hours limitations, in which all residents would have 24 + 6 hour call shifts. R1s were more divided, with 40% (18/44) favoring reversion to the pre-July 2011 system but 38% (17/44) favoring limiting all residents (including R2s/R3s) to 16 hours of continuous duty (Figure 2).
In this observational study of validated resident well-being measures at a multisite academic IM training program, we found that significantly different proportions of residents met burnout criteria and screened positive for depression when we compared survey responses between 2001, 2004, and 2012. Our residents had the lowest rate of burnout and depression in 2012, after both iterations of ACGME WHLs had been implemented. However, in contrast to this improvement in validated markers of resident well-being over time, we found that residents’ perceived impact of the 2011 WHLs on their well-being was overall negative/neutral—largely driven by the negative perceptions of R2s/R3s. Residents also perceived that the 2011 WHLs had a largely negative/neutral effect on the quality of patient care and their own medical education. Lastly, we found that most R2s/R3s and a substantial proportion of R1s indicated a preference to return to pre-2011 WHLs.
Previous single-site survey studies of resident well-being in association with the 2003 WHLs have shown an overall trend toward improvement in resident well-being after the 2003 WHLs.18–20 This notion is corroborated by the finding at our institution.14 In contrast, published studies of resident well-being markers have not shown improvement after implementation of the 2011 WHLs.11,21,22 It is difficult to say whether our study contradicts these other studies of well-being in relation to the 2011 WHLs. In our study, significantly fewer residents screened positive for depression in 2012 compared with 2004; however, there was no significant change in burnout prevalence or career satisfaction between 2012 and 2004. It is possible that the 2011 WHLs contributed to less depression at our institution; however, it is also possible that the prevalence of depression improved between 2004 and 2011 for other reasons and remained high in 2012 despite the 2011 WHLs. Importantly, in 2004, our residents perceived that the 2003 WHLs had a positive effect on their own well-being,14 while in 2012 our residents perceived that the 2011 WHLs had an overall neutral or negative effect on their own well-being. This perception supports evidence from other studies that suggests that WHL changes prior to 2011 may have improved resident well-being, while the 2011 WHLs may have had a neutral effect.
Residents at our program and in a national cohort13 felt that the 2011 WHLs had an overall neutral to negative impact on patient safety. We did not include objective measures of patient safety in our study; however, a national study of Medicare data suggested that the 2011 WHLs were not associated with a reduction in 30-day mortality or 30-day readmissions.8 The ongoing iCOMPARE study should provide another objective look at the impact of the 2011 WHL recommendations on patient care.
Even more uniform in our study was the residents’ perception of a negative effect of the WHLs on their education, with a majority of both R1s and R2s/R3s reporting a compromise to the quality of their medical education. This perception, too, is corroborated by several other studies.5,13 This makes sense given that the opportunity for residents to participate in non-patient-care educational activities, including attending rounds, self-directed reading, and teaching conferences, is in direct conflict with completing necessary patient care tasks in a timely fashion in order to leave the hospital (so-called “work compression”). Night rotations designed to achieve WHLs may also be less educational: We previously reported that night float rotations are associated with less attendance at educational conferences, less attending teaching time, and lower perceived educational value.23 Additionally, WHLs may result in proportionately more admissions that are previously admitted transfers from other teams (i.e., night float or intensive care) and already largely evaluated. Given these largely negative perceptions, it is not surprising that most R2s/R3s and many R1s favored a return to the pre-2011 WHLs structure.
Our study has several strengths, including the use of validated measures of resident well-being, the longitudinal design spanning 11 years at the same institution, and the relatively high response rate. Additionally, our program has multiple hospital settings, including county, university, and Veterans Affairs hospitals, and is typical of many residency programs.
Our study also has several limitations. Though our program includes multiple hospital settings, this was fundamentally a single-program study. Our study only provides survey results from 2001, 2004, and 2012. The long time difference between 2004 and 2012 allows significant opportunity for other confounders. For example, in our program, a new program director started in 2007. The timing of our surveys, delivered nine months after WHLs were implemented in both 2004 and 2012, allowed for perspectives from R2s/R3s who had experienced both systems, but also assessed the impact of rotation structures and workflow that had not been given time to be fully optimized. It is possible that our findings would have been different if other years were chosen. Of course, this was an observational survey study that also carries the limitations inherent to this study design. We cannot infer causality based on our findings, and the survey may have been subject to recall bias or conformity bias. Finally, because individual-respondent-level data on year in training and marital status were not available for the 2001 and 2004 cohorts, we were unable to control for potential confounding by these demographic factors. As shown in Table 1, however, demographic characteristics of residents in our program have been relatively stable over the years studied.
Overall, our study provides a unique assessment of a decade of WHLs and supports that some WHL aims with improved resident well-being may have been achieved, at least objectively, if not perceived by the residents themselves. However, we remain concerned about the potential negative impact on resident education and the uncertain impact on the quality of patient care. The ongoing iCOMPARE study may help answer some of these outstanding questions, but further research into resident workload, work compression, and sleep is essential if further WHL mandates are being considered.
The authors would like to thank the internal medicine residents and Internal Medicine Residency Program at the University of Washington and Affiliated Hospitals. They would also like to thank Dr. Lara Goitein for her help.
1. Gaba DM, Howard SK. Patient safety: Fatigue among clinicians and the safety of patients. N Engl J Med. 2002;347:12491255.
2. Shanafelt TD, Bradley KA, Wipf JE, Back AL. Burnout and self-reported patient care in an internal medicine residency program. Ann Intern Med. 2002;136:358367.
3. Weinger MB, Ancoli-Israel S. Sleep deprivation and clinical performance. JAMA. 2002;287:955957.
4. 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;351:18381848.
5. Desai SV, Feldman L, Brown L, et al. Effect of the 2011 vs 2003 duty hour regulation-compliant models on sleep duration, trainee education, and continuity of patient care among internal medicine house staff: A randomized trial. JAMA Intern Med. 2013;173:649655.
6. 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;78:11541163.
7. Reed DA, Fletcher KE, Arora VM. Systematic review: Association of shift length, protected sleep time, and night float with patient care, residents’ health, and education. Ann Intern Med. 2010;153:829842.
8. Patel MS, Volpp KG, Small DS, et al. Association of the 2011 ACGME resident duty hour reforms with mortality and readmissions among hospitalized Medicare patients. JAMA. 2014;312:23642373.
9. iCOMPARE: Comparative effectiveness of models optimizing patient safety and resident education. http://www.jhcct.org/icompare/default.asp
. Published 2015. Accessed February 7, 2017.
10. Fletcher KE, Reed DA, Arora VM. Patient safety, resident education and resident well-being following implementation of the 2003 ACGME duty hour rules. J Gen Intern Med. 2011;26:907919.
11. Ripp JA, Bellini L, Fallar R, Bazari H, Katz JT, Korenstein D. The impact of duty hours restrictions on job burnout in internal medicine residents: A three-institution comparison study. Acad Med. 2015;90:494499.
12. Mata DA, Ramos MA, Bansal N, et al. Prevalence of depression and depressive symptoms among resident physicians: A systematic review and meta-analysis. JAMA. 2015;314:23732383.
13. Drolet BC, Christopher DA, Fischer SA. Residents’ response to duty-hour regulations—A follow-up national survey. N Engl J Med. 2012;366:e35.
14. 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;165:26012606.
15. Maslach C, Jackson SE, Leiter MP. Maslach Burnout Inventory Manual. 1996.3rd ed. Palo Alto, CA: Consulting Psychologists Press.
16. Spitzer RL, Williams JB, Kroenke K, et al. Utility of a new procedure for diagnosing mental disorders in primary care. The PRIME-MD 1000 study. JAMA. 1994;272:17491756.
17. Whooley MA, Avins AL, Miranda J, Browner WS. Case-finding instruments for depression. Two questions are as good as many. J Gen Intern Med. 1997;12:439445.
18. Hutter MM, Kellogg KC, Ferguson CM, Abbott WM, Warshaw AL. The impact of the 80-hour resident workweek on surgical residents and attending surgeons. Ann Surg. 2006;243:864871.
19. Gopal R, Glasheen JJ, Miyoshi TJ, Prochazka AV. Burnout and internal medicine resident work-hour restrictions. Arch Intern Med. 2005;165:25952600.
20. Martini S, Arfken CL, Balon R. Comparison of burnout among medical residents before and after the implementation of work hours limits. Acad Psychiatry. 2006;30:352355.
21. Sen S, Kranzler HR, Didwania AK, et al. Effects of the 2011 duty hour reforms on interns and their patients: A prospective longitudinal cohort study. JAMA Intern Med. 2013;173:657662.
22. Bolster L, Rourke L. The effect of restricting residents’ duty hours on patient safety, resident well-being, and resident education: An updated systematic review. J Grad Med Educ. 2015;7:349363.
23. Luks AM, Smith CS, Robins L, Wipf JE. Resident perceptions of the educational value of night float rotations. Teach Learn Med. 2010;22:196201.