Work hours, sleep deprivation, and fatigue: A British Columbia snapshot

Sleep deprivation and fatigue and their aftereffects during medical training are topical issues in medical education. Long work hours combined with acute and chronic sleep deprivation have raised concerns about the effects of fatigue on the physician’s ability to learn, physician safety and well-being, and the quality of patient care.[1-3] Several studies have demonstrated that sleep deprivation and fatigue can adversely affect memory and cognitive performance.[4-6] More recently, it was found that residents working extended shifts in hospital are at increased risk of traffic accidents from falling asleep while driving home.[7] Furthermore, medical trainees consistently report increased levels of stress and depressed mood when sleep deprived.[8],[9] These factors lower the quality of life of medical trainees and may result in medical errors that compromise patient care.[10],[11]

In an attempt to determine the extent of sleep deprivation and fatigue in medical residents, studies in the United States have documented the work and sleep hours of medical trainees to characterize these variables in different clinical settings.[12-14] Such findings have provided a context in which to understand the scope and magnitude of these issues, as well as information for the regulation of duty hours in the United States by the Accreditation Council of Graduate Medical Education (ACGME), which limits duty hours to a maximum of 80 hours per week. In Canada, however, little is known about the work and sleep hours of medical residents. A review of the literature reveals a paucity of Canadian data exploring these issues. Furthermore, explicit guidelines for resident work-hour limits in British Columbia do not exist as they do in the United States. We therefore examined the pattern of hours worked and hours slept by first-year BC residents to determine how the experience of Canadians compares with the experience described in the published data.

Methods

University of British Columbia Faculty of Medicine postgraduate year 1 (PGY-1) residents in Lower Mainland teaching hospitals were recruited since it is a year characterized by high stress, long hours worked per week, and fewer hours of sleep per week.[13],[14] Residents were excluded if they (1) had been on vacation in the 2 weeks prior to the study, (2) suffered from a diagnosed sleep disorder, and/or (3) worked fewer than 35 hours per week (including call), which was considered part-time. A questionnaire (figure) was distributed on academic half-days in July and August 2005. Ethics approval was obtained from the UBC Behavioural Research Ethics Board.

Results

Out of 133 questionnaires distributed to first-year residents in BC Lower Mainland hospitals, a total of 104 questionnaires were returned for a response rate of 78%. Eight residents were excluded for not meeting inclusion criteria. The average age of respondents was 27.7 years (SD=2.9). Subjects represented all residency programs except for community medicine and neurology. Work hours and sleep hours were analyzed according to demographics (table 1) and rotation type (table 2). 

Residents were found to have worked an average of 65.4 hours per week, slept an average of 6.5 hours in the 24 hours preceding completion of the questionnaire, and slept an average of 41.9 hours in the previous week. Analysis revealed a wide range of work hours depending on the clinical rotation. Our results were reported based on rotation type (as opposed to residency specialty) because PGY-1 resident rotations may not accurately reflect the residency specialty of each resident’s choice. (For instance, a general surgery resident might be spending a substantial amount of first year in nonsurgery rotations such as emergency medicine and internal medicine.) Self-reported hours of work included in-house on-call activities, as residents are expected to respond to matters on the ward during any portion of this time.

Our sample reflects the national trend in medical school enrollment, with female medical trainees (55.2%) outnumbering male trainees (44.8%).[15] Females worked an average 5 fewer hours per week than male residents, yet also reported getting 2 fewer hours sleep per week than their male counterparts. These values approached statistical significance (P=.272 and P=.199, respectively). Residents who were married or had children worked fewer hours than those not married or with no children. The fewer hours of work, however, did not correspond to more hours of sleep, suggesting that sleep deprivation for this group of residents was for reasons other than work. For instance, residents may have chosen to spend this extra time meeting family and personal commitments. Our findings appear consistent with other studies that have shown male residents report longer work hours than female residents, as do single residents and residents without children when compared to married residents and residents with children.[14]

A statistical difference was observed between the surgical and nonsurgical rotations for hours worked per week, hours slept per week, and hours slept per 24 hours. These findings confirm the results of others in the literature. Baldwin and colleagues, who studied PGY-1 and PGY-2 residents in the United States, reported that the high-intensity residencies were all surgical (general surgery, obstetrics/gynecology, neurosurgery, urology, orthopaedics, and ear, nose, and throat), with an average 106.5 hours of work per week and 36.8 hours of sleep per week. These residents slept the least per week and reported the highest levels of stress and lowest levels of satisfaction among the groups.[14] In the only Canadian study of work hours reported in the literature, Lewittes and Marshall found Ontario medical residents in surgery worked the most hours per week (median 90), compared with 78 hours for internal medicine, 75 hours for family medicine, and 50 hours for psychiatry.[16] Thus, it appears that residents on surgical rotations tend to work more hours than those on nonsurgical rotations. 

Duty-hour limits

In response to concerns about the effect of long work hours on medical trainees, several countries have implemented limits on duty hours. In 2004, Great Britain and other European Union countries began adhering to the European Working Time Directive, which limits work to 58 hours a week for physicians in training (a limit that will be lowered to 48 hours per week by 2009).[17] New Zealand has mandated a 72-hour range.[18] In the United States, after much national debate, the ACGME initiated an 80-hour weekly duty limit in 2003 to prevent resident fatigue, promote well-being, and curb medical errors.[19],[20] Duty hours were defined as all clinical and administrative duties, including time spent in-house during call activities and time spent on scheduled academic activities such as conferences. The situation in Canada, however, is less clear. Work hours are regulated at the provincial level through collective agreements negotiated by resident associations, and while many agreements outline conditions such as on-call frequency limits and the number of weekends on duty, few specifically put a limit on the total number of duty hours per week including on-call activities.[21] For instance, the current Professional Association of Residents of British Columbia (PAR-BC) collective agreement regarding work hours states “A Resident shall be scheduled by the Hospital to work a reasonable number of hours. The Hospital will undertake to limit the average number of hours, having due regard for sound patient care and treatment and the educational requirements of the Resident Program.”[22] Residents of Saskatchewan and Ontario also do not have an overall duty-hour limit specified in their collective agreements, but the Professional Association of Internes and Residents of Ontario collective agreement does limit residents who do in-hospital shift work (ICU, emergency) to 60 hours per week.[23] The collective agreements for Alberta, Quebec, and the Maritimes limit residents to 12 hours of clinical duties a day; however, this limit does not include on-call hours, resulting in no clearly defined maximum.[24-26] Manitoba is the only province that clearly states a limit of 89 hours per 7-day week, including first call.[27]

Two questions remain: What is a “reasonable” number of hours for a resident to work? And should the current BC collective agreement be amended to explicitly limit the number of weekly work hours? With no clear definition of what is “reasonable” or “unreasonable” in terms of hours, it may be prudent to have more explicit wording of work-hour limits in the BC collective agreement to prevent fatigue and sleep deprivation. The ACGME has mandated the 80-hour weekly limit at all training hospitals in the United States as a starting point for addressing the issues. Since our medical training system most closely resembles the United States model, it seems reasonable to consider the ACGME’s 80-hour weekly limit as a benchmark. British Columbian PGY-1 residents on nonsurgical rotations presently do not exceed benchmarks set in the United States. BC PGY-1 residents on surgical rotations approach this limit, with two surgical rotations (general surgery at 88.9 hours per week and orthopaedics at 102.0 hours per week) exceeding the 80-hour limit. The literature suggests it is residents exceeding this limit who are at risk for traffic accidents and medical errors.[4],[6],[7],[10]

Study limitations

Our study had several limitations. Only BC PGY-1 residents were surveyed. The PGY-1 year was chosen as the literature suggests that this is the year when residents are most “at risk” because of a combination of increased work hours and inexperience. Thus, our results are less generalizable to residents in later years of training in BC or to residents in all years of training at other schools in Canada. In addition, the numbers of completed questionnaires for several specialty rotations were small (for instance, three for orthopaedics). This makes statistical validity an issue. However, our response rate of 78% for PGY-1 residents suggests our values reflect the reality of rotations at a specific time; future studies might sample groups at various times throughout the year to gain a more accurate picture overall. Another limitation is the accuracy of self-reported hours for work and sleep per week. Research suggests people may overestimate or underestimate self-reported values. While recall bias was minimized by having residents complete the questionnaire within 48 hours of receiving it, future studies might improve on the data collected by having students use a log or diary to capture the information. Furthermore, since work hours included in-house call, some residents may have been able to get some sleep while on call, thus affecting the values for hours slept without affecting total work hours. This was a limitation we had to accept since the workload variability inherent in a call night is an element beyond our control. Lastly, there were residents working at the time of our study who did not obtain and complete the questionnaire because they were not present at rounds when the questionnaire was distributed. These residents may not have attended rounds because they were too tired. Capturing information from all members of a particular PGY-1 class in the future will provide a clearer picture of hours worked and slept in a week.

Conclusions

The traditional medical school culture that sees residents working long hours and surviving on minimal sleep has slowly shifted under the mounting evidence that sleep deprivation and fatigue have detrimental effects on the physician’s ability to learn, physician safety, and the quality of patient care. Our study shows that first-year residents in BC work an average of 65.4 hours per week, which is within the recommended 80-hour weekly duty-hour limit proposed by the ACGME. However, residents on surgical rotations approach, and in some cases exceed, the 80-hour limit. With no weekly duty-hour limits in the collective agreement for residents in BC, trainees (especially those in surgical specialties) may be at risk for harm to themselves and patients because of fatigue and sleep deprivation. Further studies need to be conducted to confirm our findings in a larger sample and to correlate work hours with clinical measures of fatigue. Our study provides further insight on the topic of resident work and sleep hours from a Canadian perspective, and raises the issue that more explicit work-hour limits in BC may need to be instituted to adequately protect residents and patients from the negative effects of extended work hours.

Competing interests
None declared.

Acknowledgments
The authors wish to acknowledge the University of British Columbia Faculty of Medicine Summer Student Research Program, the Doctor, Patient and Society (DPAS 410) course, and the Canadian Institutes of Health Research, who provide salary support for Dr Miller.

Mark C. Fok, BScPharm, Andrea Townson, MD, FRCPC, Barbara Hughes, MD, FRCPC, William C. Miller, BScOT, PhD

Mr Fok is a third-year clinical clerk in the Faculty of Medicine at the University of British Columbia. Dr Townson is a clinical assistant professor in the Division of Physical Medicine and Rehabilitation, UBC, and medical manager of the SCI Rehab Program at G.F. Strong Rehab Centre. Dr Hughes is a clinical assistant professor in the Department of Medicine and Rehabilitation, and an attending physician on the SCI Program at G.F. Strong. Dr Miller is an associate professor in the School of Rehabilitation Sciences, Division of Occupational Therapy, UBC, and a researcher at G.F. Strong.