By Brad Segal
When people fall acutely ill, they deserve a non-sleep deprived doctor—but they also deserve an adequately-trained doctor. There are only so many hours to the day, and so in medical education a resident’s need for self-care must be balanced against the need for maximum clinical exposure. Since 2003, when restrictions to resident duty hours were first enacted, there has been disagreement about how to best navigate the tension. Recently, the debate resurfaced when the Accreditation Council for Graduate Medical Education (ACGME) proposed a change to the policy governing resident duty hour limits. Perhaps the most surprising part of the announcement was that their proposal increased the time limit that interns (first year residents) can care for patients without sleep. The policy ACGME enacted in 2011 had capped interns at 16 hours on-call, and the proposal increases the limit to 28 hours.
In my prior post I raised arguments for and against the proposed changes to duty hour limits. Here I will unpack the conclusions and limitations of the best empirical evidence available to ACGME: the Flexibility in Duty Hour Requirements for Surgical Trainees (FIRST) Trial. Published in the New England Journal of Medicine (NEJM) in 2016, the FIRST Trial randomized 117 surgical residency programs nationwide to have either “standard” duty hour policies, which included the current 16-hour cap on interns, or “flexible” policies, which reflect the recent ACGME proposal. Data were collected from July 2014 to June 2105. The sister-study involving medical residencies nationwide has regrettably not yet published.
The FIRST Trial warrant close attention because, like a Rorschach test, different people see different things in the data. For instance, take the finding that neither group caused significantly more or less harm to patients, though shorter duty hours were associated with more handoffs of patient responsibility. Taken at face value, these results neither clearly bolster nor contradict the proposed duty hour changes; yet they are used to both support and undermine the tentative changes to ACGME policy. The study’s first author told NPR that, “We believe the trial results say it’s safe to provide some flexibility in duty hours.” On the other hand, an editorial published in NEJM alongside the study argues that, “The FIRST Trial effectively debunks concerns that patients will suffer as a result of increased handoffs and breaks in the continuity of care.” Is there a right conclusion to draw from the study?
Let’s start by reviewing the results in closer detail. More specifically, endpoints of the study include both objective measures of patient outcomes and self-reported resident views of their duty hours.
As for objective measures, the FIRST Trial found that the change in duty hours was not associated with adverse patient outcomes. Death or serious complications occurred at the same rate among residency programs using “standard” and “flexible” duty hour policies. The two groups also did not differ when stratifying these patient outcomes by the timeliness of surgery (e.g., elective vs. emergent), operative setting (e.g., inpatient vs. ambulatory), or risk of operation (e.g., high vs. low). In secondary patient outcomes—which included development of sepsis, repeat operation, or surgical-site infection—there was again no significance difference detected between the two cohorts.
The self-reported resident outcomes in the FIRST trial yielded more between-group differences that reached statistical significance, but the results do not consistently favor one particular policy. Trainees in neither group were more likely to self-report being dissatisfied or very dissatisfied (as opposed to neutral or better) with the effect of either duty hour policy on their well-being or the quality of their education. Still, it should raise eyebrows that residents of all years with “flexible” hours were on the cusp of reaching statistical significance in their dissatisfaction about each respondent’s well-being (“flexible” group OR [95% CI] = 1.31 [0.99–1.74]; P=0.06)). The analysis did not stratify by year, so between-group comparisons for interns is not reported. While residents with “flexible” duty hours were more satisfied with continuity of care, and less frequently had to leave or miss an operation on account of duty hour limits, these residents were also significantly less satisfied with their health, rest, time with family and friends, and time for extracurricular activities.
A separate secondary analysis of the FIRST Trial data stratified resident outcomes by level of training. This offered the additional finding that interns in particular had significantly more negative views about the impact of “flexible” duty hour policies on morale, career satisfaction, and personal health. On the other hand, first-year trainees in the “flexible” group were significantly more satisfied with the quality of their education.
So far, these results support both the claim that it is safe to give residencies flexibility in scheduling, and that increased handoffs do not lead to worse patient outcomes. But understanding the limitations of the FIRST Trial are necessary for more nuanced conclusions.
First, a subsequent survey published in NEJM reported that nearly three quarters of residents had not adhered to their duty hour limits during the FIRST Trial. These violations went unreported and were not controlled for by the trial design. This undermines the extent to which this randomized controlled trial isolated the effects of duty hour limits on either patients or residents. The results should not, then, be viewed as definitive and still require subsequent validation.
Second, the FIRST Trial was entirely composed of general surgery residents. The results do not, then, establish how alterations in duty hour limits would affect other specialties. Consider, for instance, differences in culture, training, and patient care in trauma surgery vs. psychiatry.
Given that each specialty has unique considerations, the authors of the NEJM report warn against taking a one-size-fits-all approach to crafting guidelines on resident duty hours. The FIRST trial’s sister study, iCOMPARE, examines “flexible” duty hours among on internal medicine residents, but the results have yet to be made publicly available. Given the problems with extrapolating the FIRST Trial results for non-surgical residencies, it is lamentable that ACGME did not wait for the completion of iCOMPARE before taking the first step in modifying duty hour limits for trainees across all fields.
Third, the FIRST Trial used inadequate measures of Resident well-being. Validated measures of resident depression and suicidal ideation might have yielded important differences between the two duty hour policies. This is not a groundless critique, either. A 2015 meta-analysis published in JAMA found that the prevalence of depression and depressive symptoms among residents is 28.8%. A second meta-analysis did find that a comparable number of medical students are also affected by depression or depressive symptoms. Nonetheless, rates among both groups of trainees are significantly greater than age-matched peers (which nationally is 7.2% among 26- to 49-year olds, and 9.3% among 18- to 25-year olds). The causes of trainee depression are also poorly understood. Curricular requirements are believed to play some role, but the magnitude of effect is unknown due a paucity of evidence from lack of randomized controlled trials.
With these limitations in mind, then, how should ACGME’s proposal be understood in light of the FIRST trial results? First, that patient outcomes did not differ between either group objectively discredits the purported threat to patient safety from increased physician handoffs under current ACGME policies, though with the caveat that duty hour restrictions were not often enforced. Second, eliminating the restrictions on continuous hours in-house does not ensure resident satisfaction regarding their training, but it would improve aspects of education such as patient continuity, and might improve education for subsets of individuals depending the resident’s year of training or specialty. Thirdly, easing duty hour limits comes at a personal cost to all residents in terms of time with family and sleep; among interns it takes an additional toll on well-being and professional morale. It remains, though, unknown whether longer duty-hours is detrimental to the mental health of residents.
ACGME is currently sifting through input from the public before they announce their decision. Regardless of their decision, one can find reassurance in the FIRST Trial observation that neither duty hour policy harmed patients—the care that residents provide to their patients should be the overriding consideration in any proposed change to duty hours. With patient care being equal, though, it is a challenge—to say the least—to give a prescriptive formula capturing the proper trade-off between medical education and resident health. But it does not mean such a balance will not be struck eventually; it might just end up being a gradual process of randomized controlled trial and error.