This study utilized a survey instrument to evaluate the impact of duty hours' implementation on residents' career satisfaction, emotions and attitudes. In survey data, a response rate of 50% – 60% is considered acceptable for analysis and reporting . Our response rates were in that range or higher, 56% and 72%, respectively. In addition, the distribution of demographics of respondents at each survey was similar to that of the entire resident cohort at the time of the surveys. Hence, the results are representative of the entire cohort of residents of each survey time. Although there were modest improvements in these factors for the cohort after implementation of duty hour restrictions, there were notably negative changes in emotions and attitudes for selected groups, those who violated the requirements. It is likely that both the modest positive changes for the overall cohort and the notably negative changes for selected groups are both related to the impact of the restricted work hours, since almost no other changes occurred; i.e., residents' demography, salaries and benefits, clinical rotations, support staff, facilities, faculty, and institutional relationships. Between the first and second surveys, ACGME accredited programs increased from 57 to 60. Total residents (including fellows) increased from 581 to 625. Eighteen programs were primary specialties at the time of both surveys. Thirty-nine programs were subspecialty fellowships at the time of the first survey and increased to 42 at the time of the second one. Approximately 75% or trainees were in primary specialty programs at the time of both surveys.
One of the most interesting and unexpected findings of the study is related to the negative changes for those residents who violated the new requirements. Those residents held significantly more negative emotional states and were significantly more negative about the experience. It may be that those residents feel worse about their experiences for several clear reasons, most importantly because of loss of control. In many programs the onus for leaving a duty session is the responsibility of the resident, who has to make the closure of the duty period known. The individual cannot stay "till the work is done", a major change in the traditional paradigm. Not only is the resident uncomfortable giving over unfinished work to others, including the faculty, but such closure engenders negative feelings by those who are left to complete the task; i.e., faculty and senior, often chief residents. For some, leaving at a required time steals them from potentially interesting educational activities that are just beginning in which they cannot participate because their work hours have expired. Whatever the reasons, there is a significant number of trainees whose emotional states and attitudes have worsened after implementation of the work hour requirements and program directors and leaders must be aware of these changes and take action to improve them.
Fewer work hours have been reported to be associated with better attitudes and more favorable emotions. In a prospective, unpublished report comparing emotions and attitudes between U.S. and Australian PGY-1 residents over a year period, the Australian residents had many fewer negative emotions and experiences than did otherwise similar U.S. residents, and the only meaningful difference among variables was a work hour average of 56 hours for the Australians and more than 100 hours for the U.S. residents.
The significant changes in how the residency experiences are perceived are worthy of comment. The average resident after the implementation of work hours limitations reports less fatigue, more sleep, more free time and time with family. That individual also seems to feel less competent about patient care, but at the same time, less anxious about skills' development, less burdened by seeing the same kinds of patients over and over and less criticized by staff. All these positive changes are consistent with the implementation of reduction in work hours.
Several limitations to the study could be cited. One perceived limitation is that this study is a report from a single institution, albeit a relatively large one. We do not believe that this fact lessens the generalizability of the results, since virtually all previously reported studies from this institution have shown consistent patterns [1
], even when cohorts from other institutions were studied simultaneously. Self-reporting of duty hours has some inherent validity concerns. However, the vast majority of studies of similar themes and reports from the ACGME have used self reported data [18, 19, 20, 23, 24], and our institutional data compare favorably with the results of ACGME residents' surveys and ACGME site visit outcomes. Because the requirement did not exist at that time of the first survey, it contained no questions about duty hours. Thus, direct comparison of work hours between the two periods is not available. However the Division of Graduate Medical Education has information from its internal reviews, electronic monitoring systems, ACGME resident surveys and site visits about all programs' changes in duty hours over the study period. In July 2003, with initial implementation of the duty hour requirements, all 54 ACGME accredited programs were surveyed to identify program frequencies of non-compliance with the requirement for 80 hours per week averaged over a 4-week period. Reporting rates ranged from 18–100%. From that information compliance data were calculated. Twenty-two programs, representing approximately half of all residents, were considered at high risk for non-compliance because one or more resident reported excess working hours. Between July 2003 and the time of the December 2004 survey, reports of violations were significantly reduced although not eliminated. At that time six programs, representing approximately 33% of all residents, continued to be out of compliance by at least one respondent.
We believe that work hour relief is important and should be a contribution to improved emotions and perceptions of experiences with residency training. And, for most residents that seems to be the case. In those programs where implementation of these requirements results in more negative changes, program directors and department leaders must innovate. Part of that process must include that the faculty fully support the residents' departure at the appropriate time, that there be appropriate resources to "transition" patients' care, and that there be alternative experiences for those residents who "miss out on opportunities" because of work hour limits. The important lesson from this study is that work hour limits seem to improve emotional and attitudinal factors for most, but not all residents. For the latter groups, innovation is required to improve the experiences.