Included is a section: “What Have We Learned About Limiting Physician Work Hours?”

Recent Reviews and Systematic Evaluations

• Multiple systematic reviews have addressed the patient safety effects of reducing shift length for residents. One systematic review that included only studies of the effect of the 2003 ACGME regulations identified 20 studies that assessed mortality and 24 studies that assessed other patient safety outcomes before and after implementation of the regulations. Meta-analysis of the mortality studies did show a statistically significant decline in mortality after 2003 (OR 0.9, 95% CI 0.84 – 0.95), which was consistent in studies examining either medical or surgical patient populations. However, considerable unexplained heterogeneity was present (I2=83%), and the subset of studies that used a contemporaneous control group of non-teaching hospitals generally did not find a difference in mortality.

The authors acknowledged that as they were unable to control for secular trends or changes in patient characteristics, the mortality improvement could be due to overall improvement in the quality of care during the time period studied. The studies examining patient safety outcomes yielded mixed results, with no clear pattern of improvement or worsening across studies.

• Another review identified 36 studies that examined the association between reduced trainees’ working hours and patient outcomes. This study also included studies performed outside the U.S.. The conclusions were largely similar to that of the previous review:

Mortality and patient safety outcomes appeared unchanged after implementation of duty hour limits. Both reviews found that limiting work hours appeared to improve residents’ quality of life.

• The question of why patient safety outcomes have not improved after reducing resident work hours is a subject of intense debate. Reduced shift length almost certainly led to greater discontinuity among providers, and the resultant handoffs of care may have had deleterious effects on patient safety. Adherence to work hour limitations was (and is) likely suboptimal.

 In addition, studies have shown that changing residents’ work schedules to meet the regulations did not actually result in residents sleeping more. Finally, although resident quality of life improved, in some studies objective measurements of burnout and depression among residents did not change.

 Burnout and depression are themselves linked to impaired job performance, independent of acute or chronic fatigue.

• The 2003 duty hour regulations still allowed all residents to work a maximum of 30 consecutive hours. These extended duration shifts are still longer than those allowed in virtually any other industry, and studies have found an association between working more than 16 consecutive hours and an increased risk of self-reported errors and attentional failures.

• A 2010 systematic review that identified 13 studies in which shift length for clinicians was purposefully reduced found consistent evidence among the higher-quality studies that both objectively-measured and self-reported errors decreased after shift length reduction.

One particularly high quality study found a significant reduction in serious medical errors for medical interns assigned to work a 16-hour shift in the intensive care unit, compared with interns working a traditional 30-plus hour shift. However, the reviewers were unable to reach a firm conclusion regarding the optimal shift length, due to heterogeneity between shift lengths used in the primary literature.

• Thus, the totality of the evidence on the 2003 ACGME duty hour regulations indicates that reducing resident duty hours does not improve—or worsens—patient safety or mortality. The association between extended duration (>16 hour) shifts and adverse events ultimately was a factor in the ACGME’s decision to enact a 16-hour shift length limit for first-year residents as part of the 2011 regulations.

Potential for Harm

• The greatest potential harm of work hour regulations is an increase in adverse events due to increased handoffs of care between providers. Although this association is certainly plausible, and handoffs have unquestionably increased after both the 2003 and 2011 regulations, studies have not specifically examined whether errors attributable to handoffs have increased after the regulations were implemented.

• The other oft-cited adverse consequence of duty hour reduction is decreased clinical experience for trainees, limiting their ability to practice independently once training is completed. Studies of the 2003 duty hour regulations generally did not find that objective clinical experience worsened, when measured by criteria such as surgical case volume. However, both faculty and residents have voiced concerns that duty hour regulations have actually compromised their educational experience, and most residents appear unconvinced that further duty hour reductions will improve either patient safety or their educational experience.

Costs and Implementation

• Implementing the 2011 ACGME regulations is likely to be extremely costly for teaching hospitals. A 2011 cost-effectiveness analysis estimated that implementing the new ACGME regulations would cost teaching hospitals $1.6 billion if the decreased workload of interns was replaced entirely by attending physicians, and $1.34 billion if interns were replaced by physician extenders (nurse practitioners or physician assistants). A 7.2% decrease in preventable adverse events would be required in order to make the regulations cost-neutral to society, but teaching hospitals would still encounter considerable costs. The expenses associated with the need to replace the housestaff workforce with alternative providers and the need to provide greater

Conclusions and Comment

• Sleep deprivation and fatigue have clear deleterious consequences for patients and providers.

However, the most prominent effort to improve patient safety by reducing fatigue—limiting the work hours of resident physicians—has not yielded the expected benefits.

It is conceivable that the 2003 ACGME duty hour regulations were simply not stringent enough, given that extended duration shifts were still permitted and those shifts are associated with preventable adverse events.

Alternatively, it may be that advocates underestimated the complexity of the relationship between duty hours and safety, or the detrimental impact of handoffs.

The 2011 regulations further restrict hours, particularly for first-year residents. The effects of these new regulations are as yet unknown, and unfortunately, the existing evidence does not offer us great clarity regarding the optimal work hour structure that would improve safety by decreasing clinician fatigue with minimal potential for unintended consequences