Secondary Logo

Journal Logo

Duty Hours

Impact of Reduced Duty Hours on Residents' Educational Satisfaction at the University of California, San Francisco

Vidyarthi, Arpana R. MD; Katz, Patricia P. PhD; Wall, Susan D. MD; Wachter, Robert M. MD; Auerbach, Andrew D. MD, MPH

Author Information
  • Free


In July 2003, the Accreditation Council for Graduate Medical Education (ACGME) limited residents' duty hours to 80 hours per week, with one day off per week and no more than 30 consecutive work hours.1 The decrease in duty hours affects 8,000 ACGME-accredited training programs and their 100,000 trainees, not to mention the 16.8 million patients admitted to teaching hospitals each year.2 The ACGME mandate was instituted partly in response to growing concerns about the impact of long duty hours on residents' clinical and educational performance3 and is being actively enforced.4 Implementing duty-hour limitations requires training programs and hospitals to make large commitments in terms of administrative time, personnel, and dollars directed at designing new systems and enforcing the mandates while maintaining quality of training programs and patient care.5

There is surprisingly little evidence on the impact of reduced duty hours on residents' education, and the few data that exist suggest reducing duty hours produces mixed results. In one study, surgical residents were split in their perceptions of overall educational impact after duty-hour reductions.6 Another suggested that duty-hour reduction had a mixed impact on internal medicine residents' intellectual interest in challenging medical problems.7

To examine the impact of duty-hour reduction on educational satisfaction of internal medicine residents, we surveyed residents in a large university-affiliated training program after system changes were implemented to reduce duty hours. This cross-sectional study aims to describe residents' perceptions of the value of particular educational activities, identify which administrative/clerical tasks interfere with those educational activities, and evaluate the impacts of duty-hour reduction on those activities and tasks and on overall educational satisfaction. We hypothesized that educational satisfaction after duty hours were reduced would not improve given the current administrative demands on residents' time.


Sites and participants

We conducted our study at the University of California, San Francisco (UCSF), School of Medicine internal medicine training program, which supports residency training at three clinical sites: the San Francisco Veterans Affairs Medical Center, San Francisco General Hospital (a municipal teaching hospital), and Moffitt-Long Hospital (an academic medical center). All internal medicine residents at UCSF were eligible to participate in the study. Internal medicine residents at UCSF are in three programs: preliminary, one year of internal medicine for residents transitioning into another specialty such as radiology or dermatology; primary care, a three-year residency focused on ambulatory care; and categorical, a three-year residency with broad foci. The study was approved by UCSF's institutional review board.

In February 2003, each of the three UCSF-affiliated hospitals implemented system changes designed to reduce duty hours and comply with the ACGME mandates. Prior data collected from each site showed that residents were routinely working more than 80 hours per week. These system changes, designed with residents' and faculty's input, varied slightly from site to site, but included the use of daytime cross-coverage residents (“day float”) to provide coverage to postcall teams (thereby facilitating early departure of the postcall team), overnight cross-coverage (“night floats”) to relieve the on-call team from early morning admissions and provide coverage of noncall teams, and early sign-out times facilitated and enforced by the chief residents at each site.

Questionnaire development and administration

We developed the questionnaire using a three-step process. First, three authors (ARV, PPK, ADA) identified salient domains relevant to residents' educational experience and workload based on one month of observing residents working on the inpatient service at Moffitt-Long Hospital. The principal investigator (ARV) developed preliminary questions based on these domains and review of the existing literature. To establish content validity, these questions were posed to experts in the fields of medical education, outcomes research, and psychometrics at UCSF, after which items were reworded, reformatted, or removed entirely. In addition, new questions were added based on expert opinion. Finally, we pretested the questionnaire in a sample of noninternal-medicine housestaff at UCSF and recent residency graduates outside of UCSF as a final check for usability and clarity. In addition to collecting basic demographic data, our questionnaire had questions in four broad domains:

  • Residents' perceptions of the value of educational activities considered core to residency training, rated on an 11-point scale (0 = no perceived value, 10 = highest perceived value)
  • Residents' perceptions of work tasks and overall workload, particularly how frequently administrative tasks interfered with learning, rated on a five-point scale (1 = never, 5 = very often)
  • Questions regarding residents' satisfaction with their education and work, rated on 11-point scale (0 = completely unsatisfied, 10 = completely satisfied)
  • Residents' perceptions of how education and work may have changed after duty hours were limited, specifically the amount of time spent in activities after duty-hour reduction, rated on a five-point scale (1 = less time, 5 = more time), and the perceived impacts of the new system with the following responses: negative, no impact, and positive impact

We collected questionnaires from internal medicine residents at UCSF for four months after duty hours were reduced starting in March 2003, one month after implementation of the system change to reduce duty hours until June 2003, to ensure residents had time to experience the system changes. During this time, 164 residents were eligible to participate. To maximize the number of respondents, we mailed a questionnaire to residents' home addresses with a US $3 financial incentive and sent multiple follow-up reminders to nonrespondents via e-mail, postal mail, and in conferences. Questionnaires were randomly coded to track response rates and these codes were removed prior to data entry to ensure confidentiality. Data were entered by a professional vendor and double-keyed to ensure accuracy with no errors detected. Participation was voluntary and consent was implied with return of the questionnaire.

Statistical analysis

We used univariate statistics to characterize the distribution of residents' responses; repeated measures analysis of variance (ANOVA) were then used to determine the statistical significance of differences between mean ratings of educational activities and interfering administrative tasks.

We tested four items intended to measure work satisfaction (“During your most recent inpatient rotation work week, how satisfied were you with the following? A. workload, B. work life, C. quality of patient care provided and D. quality of teaching provided”) using Spearman correlation analyses. These analyses confirmed responses to the four items were highly correlated allowing us to combine them to form a single work satisfaction summary score. This score was calculated as the mean of each resident's responses to the four items, and yielded a scale ranging from 0 to 10. The scale appeared to be reliable with internal consistency (Cronbach's alpha 0.86). Our measure of educational satisfaction was a single item reported on a 0–10 scale.

Multivariate linear regression models were used to identify factors associated with both the work satisfaction summary score and the educational satisfaction score. We selected variables for entry into the regression models based on observed bivariate relationships with the dependent variable and confounding with other independent variables. Covariates in multivariable models included: age (>30 years versus ≤ 30 years), sex, postgraduate year, type of program (primary care versus categorical), perceived percentage of time spent on administrative tasks (non-MD tasks: less than 50% versus greater than 50%), feeling overwhelmed at work, and whether the individual reported working more than 80 hours per week even after work-hour limitations were implemented. We performed all analyses using SAS version 8.12 (SAS Institute Inc., Cary, NC.).


A total of 125 residents (76%) returned completed questionnaires. Comparing demographic data known on all residents, sex (p > .2), level of training (p > .45), and type of program (p > .6) were similar between respondents and nonrespondents. Respondents were equally distributed among year of training (PGY-1, 36.6%; PGY-2, 35.8%; PGY-3, 27.6%). Most respondents were female (60%), enrolled in the categorical residency track (62%), and younger than 30 years of age (70%) (Table 1). All residents stated that they were aware of the system changes intended to reduce duty hours to fewer than 80 per week and 92% (n = 115) of residents responded they believed that residents should work fewer than 80 hours a week. At the time of the study, all residents had experienced an inpatient ward month prior to and after the system changes. Thirty-five percent (n = 44) stated that they had worked more than 80 hours per week during their most recent inpatient rotation after the system changes for reducing duty hours.

Table 1
Table 1:
Characteristics of Residents Who Responded to a Questionnaire on the Impact of Reduced Duty Hours on Educational Satisfaction, Three Clinical Sites, University of California, San Francisco, School of Medicine, 2003

Perceived value of educational activities

Residents' ratings of educational activities are shown in Figure 1. Using ANOVA techniques to determine statistically different groups of activities, we were able to define groups of activities that were rated highly, those with intermediate ratings, and those rated least positively.

Figure 1
Figure 1:
Residents' mean ratings of various educational activities (0 = no value, 10 = highest value), from a study on the impact of reduced duty hours on educational satisfaction, at three clinical training sites affiliated with the University of California, San Francisco, School of Medicine, 2003. Vertical error bars indicate the standard deviation, with horizontal arrows representing p values for comparisons between individual responses or groups of responses. Group 1: morning report and teaching others; the mean value difference was not significant. Group 2: attending rounds, reading, and noon conference; the mean value difference was not significant. Group 3: journal club and grand rounds; the mean value difference was not significant. Group 1 versus Group 2: the mean value difference was statistically significant. Group 2 versus Group 3: the mean value difference was statistically significant.

Morning report and teaching others (e.g., residents and students) were rated most highly (mean 8.2, SD 1.8; mean 8.1, SD 1.6, respectively). Ratings of these activities did not differ from each other (p = .978), and as a group differed from all other activities (p < .0001 compared to the intermediate and lowest rating group). Attending rounds (mean 7.23, SD 1.5), reading (mean 6.91, SD 2.3), and daily noon conference (mean 6.84, SD 1.6) rated similarly, and formed an intermediate rating group. The intermediate group rated higher than did the lowest rating group which included journal club (mean 5.1, SD 2.5) and grand rounds (mean 4.8, SD 1.8, p < .0001).

Reports of administrative tasks that interfere with resident education

Answering pages most frequently interfered with residents' ability to engage in educational activities as compared to all other potentially interfering activities (mean 4.4 on a five-point scale, SD 0.69, p < .0019 compared paperwork and waiting on hold, and p < .0001 to all others). Paperwork and waiting on hold on the telephone were the next most common interfering activities, (mean 4.2, SD 0.90; mean 4.1, SD 0.967, respectively), not significantly different from each other (p = .258) but with higher interference scores than scheduling tests (mean 3.8, SD 1.0) and appointments (mean 3.6, SD 1.1, p < .0001) (see Table 2).

Table 2
Table 2:
Administrative Tasks that Residents Reported Interfered with Educational Activities, Three Clinical Sites, University of California, San Francisco, School of Medicine, 2003

Impact of duty-hour reduction

Approximately two thirds of residents reported they spent the same time or less time in highly rated educational activities such as teaching others and in conference, and more than one quarter of residents stated that their ability to attend conferences was negatively affected after duty hours were reduced. Most residents (60 %) reported that duty-hour reduction did not increase the amount of time they spent reading. Of note, though, approximately one third spent more time reading and teaching others after duty hours were reduced. The amount of time residents spent on tasks that they reported interfered with their educational activities, such scheduling appointments, did not change after duty-hour reduction (51% unchanged) (see Table 3).

Table 3
Table 3:
Residents' Reported Impact of Reduced Duty Hours on Daily Educational and Administrative Activities, Three Clinical Sites, University of California, San Francisco, School of Medicine, 2003

When asked about the impact of duty-hour reduction on aspects of resident life, the vast majority of residents stated that the reduction had a positive impact on quality of life (78%) and hours worked (82%), but the majority (72%) reported that amount of time spent doing non-MD oriented tasks was unchanged. Although residents' reports on how duty-hour restrictions affected education overall were split, 28% reported a negative impact on education which was the only aspect of resident life queried with a substantial number of reports of a negative impact due to reduced duty hours (see Table 4).

Table 4
Table 4:
Residents' Reported Impact of Reduced Duty Hours on Aspects of Resident Life, Three Clinical Sites, University of California, San Francisco, School of Medicine, 2003

Predictors of overall satisfaction versus educational satisfaction

The mean overall work satisfaction summary score for residents was 6.5 on a 0–10 scale (SD 1.7). The mean satisfaction with educational experience was 7.0 (SD 1.8).

In multivariable models (including sex, age, PGY, type of program, the amount of time spent on administrative tasks, perceived working more than 80 hours a week, and feeling overwhelmed at work), factors independently associated with lower overall work satisfaction were PGY-1 (1.1 points less satisfied, p = .004), feeling overwhelmed at work (0.8 points less satisfied, p < .0001), and working more than 80 hours per week (0.6 points less satisfied, p < .05). In the multivariable model examining predictors of educational satisfaction, including the same variables as the previous model, only PGY-1 and feeling overwhelmed with work were significantly associated with lower educational satisfaction. (0.98 points less satisfied, p < .05, and 0.55 points less satisfied, p = .0111 respectively) (see Table 5).

Table 5
Table 5:
Predictors of Residents' Educational and Work Satisfaction after Duty-Hour Reduction, Three Clinical Sites, University of California, San Francisco, School of Medicine, 2003


In this cross-sectional survey of internal medicine residents, we found that duty-hour reduction was associated with improved ratings of quality of life, but not with reports of increased time spent in educational activities or improved educational satisfaction.

Residents in our study, as in other similar studies, were more satisfied with their quality of life after duty-hour reduction.6 Given the high rates of depression and burnout for residents,8,9–10 improved quality of life is, in itself, an important outcome of duty-hour changes and should not be underemphasized. However, our data suggest that improved quality of life was not accompanied by improvements in educational satisfaction or work-life ratings.

Although some residents did find more time to engage in teaching and reading, overall, residents in our study were not more satisfied with their education after duty hours were reduced. As residents are adult learners and consumers of education, their satisfaction may be a marker of overall quality of education, just as patient satisfaction is a factor in measuring overall health care quality.11 One of the hopes of the ACGME in mandating duty-hour reduction was that shorter hours would make residents better able to learn.3 Our data, however, show that, from the perspective of the residents themselves, this educational goal may not have been realized by a decrease in duty hours alone.

The disconnect between quality of life and other outcomes in our study suggests that competition between education and administrative tasks needs to be considered when implementing duty-hour reduction. A possible mechanism of the disconnect is that residents in our program perceived no change in administrative or clerical work after duty hours were reduced. Residents spend a large percentage of their time doing paperwork and procedures that might be done by nonphysicians,12 activities generally considered of low educational value.13 Duty-hour reductions may, in fact, simply compress the same amount of low-value work into a smaller period of time, exacerbating residents' need to skip educational activities as we have described, or magnify the onerousness of administrative tasks. Our observation that PGY-1 residents, the trainees most burdened with administrative work, were less satisfied with their work and education helps confirm this connection.

Although our data do not allow us to explore all the potential mechanisms and effects of duty-hour reductions, they do provide tantalizing clues to how quality of life improvements may not improve educational experience. For example, if residents skip conferences or work during teaching sessions (i.e., writing a progress note while listening to a lecture) to leave the hospital at a fixed time, residents may be more rested and feel generally improved although their educational experience is lessened.

Interestingly, residents who claimed that they continued to work more than 80 hours a week after the initial duty hours changes, although less satisfied with work life, were not less satisfied with education. Although we cannot explain these findings directly with our data, in concert with our findings regarding clerical tasks, they reinforce the idea that the number of hours that residents work is not directly linked to educational satisfaction.

Our study had several limitations. Our results are from a university-based training program and may not be directly applicable to training programs in private or community-based hospitals. Because we only surveyed internal medicine residents, our findings may not extend to other training programs with differing educational requirements (e.g., surgery training programs). As a single-system study, our results may have limited applicability to other settings. However, we assessed the impact of a variety of “day float” and “night float” interventions implemented at our three hospitals. We believe that approaches chosen at UCSF sites are very similar to those being attempted elsewhere.14–16 As a retrospective survey, residents' recollections may be subject to recall bias. Our survey methods sought to minimize this bias by timing the questionnaire so it was administered fairly soon after the first month of duty-hour reductions, and by asking questions that explicitly framed the comparison we hoped to assess. Although we did not validate residents' perceptions with corroborating workload data, our findings were generally consistent with existing time-motion studies.12,17

Our results suggest that ACGME duty-hour requirements appear likely to meet one of their goals by improving the overall well-being of residents. A positive impact on resident education may not be as likely, however, as quality of life may be improved at the expense of education. To ensure that work hour reduction does not come at the cost of educational satisfaction, hospitals and residency programs should consider methods for optimizing support of the residents, particularly those who are stressed for time due to inexperience or high administrative burden, such as PGY-1 residents. These interventions, in consort with reduced duty hours, may help to meet the goal of an improved educational experience in the era of duty-hour reduction.

Dr. Auerbach is supported by a Mentored Research Career Development Training Grant from the Agency for Health Research and Quality (K08 HS11416–02). The authors gratefully acknowledge the following for their assistance in conception and compilation of this study: Marla Eisenberg, ScD, Susan Nguyen, and David Irby, PhD


1 Accreditation Council for Graduate Medical Education. Resident Duty Hours Common Program Requirements 〈〉. Accessed 21 September 2005.
2 Health Care Cost and Utilization Project. Agency for Healthcare Quality and Research 〈〉. Accessed 21 September 2005.
3 Philibert I, Friedmann P, Williams WT. New requirements for resident duty hours. JAMA. 2002;288:1112–14.
4 Croasdale M. Johns Hopkins penalized for resident hour violations: internal medicine residency is threatened with loss of accreditation 〈〉. Accessed 20 September 2005.
5 Vogt K. Hospitals count up cost of reduced resident hours: facilities have hired new staff or shuffled schedules to make up for time lost to the 80-hour workweek limit 〈〉. Accessed 28 September 2005.
6 Barden CB, Specht MC, McCarter MD, Daly JM, Fahey TJ, 3rd. Effects of limited work hours on surgical training. J Am Coll Surg. 2002;195:531–38.
7 Conigliaro J, Frishman WH, Lazar EJ, Croen L. Internal medicine housestaff and attending physician perceptions of the impact of the New York State Section 405 regulations on working conditions and supervision of residents in two training programs. J Gen Intern Med. 1993;8:502–7.
8 Bellini LM, Baime M, Shea JA. Variation of mood and empathy during internship. JAMA. 2002;287:3143–46.
9 Shanafelt TD, Bradley KA, Wipf JE, Back AL. Burnout and self-reported patient care in an internal medicine residency program. Ann Intern Med. 2002;136:358–67.
10 Thomas NK. Resident Burnout. JAMA. 2004;292:2880–89.
11 Wachter RM, Goldman L, Hollander H. Hospital Medicine. Philadelphia: Lippincott Williams & Wilkins, 2000.
12 Lurie N, Rank B, Parenti C, Woolley T, Snoke W. How do house officers spend their nights? A time study of internal medicine house staff on call. N Engl J Med. 1989;320:1673–77.
13 Dresselhaus TR, Luck J, Wright BC, Spragg RG, Lee ML, Bozzette SA. Analyzing the time and value of housestaff inpatient work. J Gen Intern Med. 1998;13:534–40.
14 Buff DD, Shabti R. The night float system of resident on call: what do the nurses think? J Gen Intern Med. 1995;10:400–402.
15 Carey JC, Fishburne JI. A method to limit working hours and reduce sleep deprivation in an obstetrics and gynecology residency program. Obstet Gynecol. 1989;74:668–72.
16 Lieu TA, Forrest CB, Blum NJ, Cornfeld D, Polin RA. Effects of a night-float system on resident activities and parent satisfaction. Am J Dis Child. 1992;146:307–10.
17 Moore SS, Nettleman MD, Beyer S, et al. How residents spend their nights on call. Acad Med. 2000;75:1021–24.

Did You Know?

The nation's first cardiac clinic was established in 1919 at the University of Pennsylvania School of Medicine.

For other important milestones in medical knowledge and practice credited to academic medical centers, visit the “Discoveries and Innovations in Patient Care and Research Database” at 〈〉.

© 2006 Association of American Medical Colleges