Longer work hours among resident physicians were significantly associated with increased stress and higher self-perceived clinical competency, but not with burnout, according to a cross-sectional analysis of 540 US residents in high-burnout specialties. The findings suggest that the 2003 duty-hour restrictions from the Accreditation Council for Graduate Medical Education may have limited utility in addressing the burnout epidemic that continues to affect graduate medical education.
The study found no statistically significant association between burnout scores and average weekly work hours or hours worked in the prior week in linear regression models controlling for demographic variables including gender, race, ethnicity, specialty, training level, relationship status, geography, and season of survey completion, wrote lead study author Sydney F. Tan, MD, of the University of Wisconsin School of Medicine and Public Health, and colleagues in JAMA Network Open.
In contrast, stress showed robust positive associations with both work-hour measures. Using the 10-item Perceived Stress Scale (PSS), researchers observed that stress was significantly associated with average hours worked and hours worked the prior week. In unstandardized terms, each 10-hour increase in weekly work hours corresponded to approximately 1-point increases on the PSS (range, 0-40; higher scores indicate greater perceived stress).
Perhaps more consequential for ongoing debates about duty-hour policy, self-assessed Accreditation Council for Graduate Medical Education (ACGME) competency milestones were also positively associated with work hours. For every 10-hour increase in average weekly work, milestone scores increased by 0.11 on the 5-point scale. The researchers noted this finding "adds to the body of evidence that work hour restrictions may have negative consequences on resident education and self-efficacy."
The survey was conducted from February through June 2024 as part of baseline assessment in a randomized clinical trial evaluating a meditation-based well-being intervention. Eligible participants were residents in general surgery and surgical subspecialties, ObGyn, family medicine, internal medicine, and emergency medicine. The mean average weekly hours worked was 65, with mean hours worked in the prior week at 60.
Participants skewed female compared with national ACGME resident data: 67% identified as cisgender women vs 47% in ACGME 2022-2023 data for the same specialties. Racially, 67% identified as White, 21% as Asian, and 5% as Black. Medical specialties comprised 56% of the sample, with surgical specialties at 44%.
Burnout was measured using the 9-item abbreviated Maslach Burnout Inventory-Human Services Survey depersonalization and emotional exhaustion subscales (range, 0-36; higher scores indicating higher burnout). Personal accomplishment was assessed separately using the personal accomplishment subscale (range, 0-18; higher scores indicating higher personal accomplishment and lower burnout). Neither burnout nor personal accomplishment showed statistically significant associations with work hours.
The researchers explored a comprehensive set of potential moderators including resilience, sleep disturbance, loneliness, flourishing, meaning and purpose, and mindfulness facets. Resilience appeared to moderate the work hours-stress relationship in unadjusted analysis, suggesting those with higher resilience were less susceptible to the stress effects of longer hours. However, this finding did not survive correction for multiple comparisons. The researchers noted that "despite investigating a large set of candidate moderators, we found little evidence that demographic or well-being characteristics moderate the association between work hours and outcomes."
The dissociation between stress and burnout presents a nuanced picture. As the researchers explained: "While work hours alone may not explain burnout, the association of work hours with elevated stress remains important as prolonged exposure to overwhelming stress leads to burnout." They suggested that while stress represents a situational response to immediate demands, burnout constitutes a "pathologic state of prolonged, intense stress" involving exhaustion, cynicism, and reduced accomplishment.
The findings have direct policy implications. "Our findings challenge the notion that further restricting work hours alone will reduce burnout," the researchers wrote, "and suggest further restrictions could negatively impact residents' perceived competency." However, they cautioned that results "should not be used to support arguments to eliminate work-hour restrictions, either." Instead, they argued for more comprehensive approaches addressing "what occurs during work hours, life outside of work, an individual's relationship to work, and the organizational culture and broader systems within which residents work."
Commentary: Understanding Stress vs Burnout
In an accompanying invited commentary, Areti Tillou, MD, and Jonathan R. Hiatt, MD, of the David Geffen School of Medicine at University of California, Los Angeles, emphasized the importance of distinguishing between stress and burnout when interpreting these findings. "While stress is situational, burnout is an emotional, physical and mental state," they wrote. "Understanding the distinction between the [two] concepts is important in the interpretation of the results and potential solutions."
The commentators highlighted the concept of eustress, introduced by Hans Selye in 1936s, where stressors are perceived as positive challenges rather than threats. "Eustress might be necessary to maximize performance, while higher levels of stress lead to exhaustion and burnout," they noted, adding that "complete elimination of work-related stress is neither possible nor desirable, as the presence of eustress may be of benefit."
Drs. Tillou and Hiatt raised methodological concerns about the heterogeneity of work hours across specialties. They cited prior research showing general surgery, internal medicine, and anesthesia/radiology residents reported approximately 84, 69, and 52 mean weekly hours, respectively, with no effect of total work hours on burnout or well-being. They also noted that 75% of study participants were midlevel trainees or lower, and that burnout fluctuates over training, increasing in the first 3 years before declining sharply according to 2024 American Medical Association data.
The commentators cited a meta-analysis of 48 studies supporting the conclusion that burnout is influenced by more than hours worked. That analysis found trainees reporting negative workplace conditions, such as dysfunctional work environment, excessive work demands, and concerns about patient care, were twice as likely to report burnout or stress. Nonwork-related factors, while contributing, were less impactful.
Despite the limitations, the commentators called the study "a useful contribution to the topic" and suggested the well-being intervention being evaluated in the parent trial "may provide constructive tools for next steps and help training programs to find a balance that mitigates burnout and optimizes well-being and competency development for their residents."
The cross-sectional design limits causal inference, and the researchers acknowledged potential selection bias from recruiting through a well-being intervention trial. Response rate could not be calculated due to electronic and snowball recruitment methods. All measures were self-reported; program evaluation data were not used to preserve anonymity. The competency milestone findings reflect resident self-assessment rather than faculty clinical competency committee evaluations.
The researchers concluded that "resident burnout is influenced by more than just number of hours worked" and called for research identifying "interventions that support both resident well-being and competency development" in contemporary training environments that have evolved alongside policy, technological, and workforce demographic changes since duty-hour restrictions were first implemented. The full study and commentary are available in JAMA Network Open.
Disclosures can be found in the study; the authors of the invited commentary declared having no competing interests.
Source: Study, Invited Commentary
