Intensive Care Unit Resident Scheduling Trial

Background: Strategies to manage residents' fatigue must balance patient safety, resident education and resident wellbeing. The trade‐offs among these are not fully understood. A Canadian cluster‐randomized clinical trial will provide urgently needed evidence to inform resident scheduling practice and policy. Previous work questions assumptions used to justify duty hour reduction. The investigators found residents working overnight are fatigued but do sleep, are not chronically sleep deprived, learn effectively immediately after being 'on‐call' and learn in a 4‐week ICU rotation. The pilot cluster randomized clinical trial of 12, 16 and 24h overnight duty suggests that patient safety is compromised with duty periods shorter than 24h: more harmful errors occurred in the 12h schedule, and residents' knowledge of patients and clinical decisions were worst in the 16h schedule. Mortality was similar. Resident wellbeing was worst in the 24h schedule, suggesting a trade‐off between patient safety and resident wellbeing. Education was not assessed. The 8 other randomized clinical trials of physician schedules are from the US; 2 studied residents (the providers of first line overnight medical care in Canadian hospitals), and 6 had low power for important effects on patient outcomes. None found differences in mortality or harmful errors or robustly examined educational outcomes. With Canadian Institutes of Health Research bridge funding the investigators completed a Canada‐wide survey showing that most ICU overnight in‐house physician staffing is by residents, and a pilot of education outcomes demonstrating the feasibility, responsiveness and discriminative power of competency assessment. Goals: To evaluate the effects of 16h and 24h resident duty schedules on patient mortality and safety, resident education and resident wellbeing. Design: A cluster‐randomized crossover trial will compare 16h vs. 24h overnight schedules for residents rotating to ICU. Eligible ICUs will care for adult patients, and are anticipated to have rotating residents performing overnight in‐house duty. Intervention: 16h and 24h overnight schedules will both be applied for 52 weeks at each site. Schedule crossover order will be randomly allocated (1:1 ratio). All in‐house residents will participate in the schedule. Consent will be obtained for resident measurements. Outcomes: are in 3 domains, Patient, Resident Education and Resident Wellbeing. The primary outcome is hospital mortality to 90 days following index ICU admission. Mortality is objective, patient‐relevant, frequent (12‐20% in adult ICU) and reflects the quality and safety of care. The main resident education outcome is cognitive reasoning, and the main resident wellbeing outcome is emotional exhaustion. Study of patients and residents in 18 ICUs has power >90% for a 2% difference in mortality and of 90% for important differences in resident education and resident wellbeing. Analyses will use hierarchical regression models to account for clustering by ICU. Expertise: The research team includes experts in patient safety, postgraduate medical education, randomized clinical trials, sleep, and 2 national‐level decision‐makers. Impact: Key stakeholders actively seek high‐quality data about the effects of common overnight schedules on patients and residents. Understanding the benefits and trade‐offs will support creation of evidence informed policy about resident schedules and mitigation strategies. This knowledge will improve care for patients and help better train doctors.