Intensivists' Direct Management without Residents May Improve the Survival Rate Compared to High-Intensity Intensivist Staffing in Academic Intensive Care Units: Retrospective and Crossover Study Design

摘要

Background Medical staff members are concentrated in the intensive care unit (ICU), and medical residents are essentially needed to operate the ICU. However, the recent trend has been to restrict resident working hours. This restriction may lead to a shortage of ICU staff, and there is a chance that regional academic hospitals will face running ICUs without residents in the near future. Methods We performed a retrospective observational study (intensivist crossover design) of medical patients who were transferred to two ICUs from general wards between September 2017 and February 2019 at one academic hospital. We compared the ICU outcomes according to the ICU type (ICU with resident management under high-intensity intensivist staffing vs. ICU with direct management by intensivists without residents). Results Of 314 enrolled patients, 70 were primarily managed by residents, and 244 were directly managed by intensivists. The latter patients showed better ICU mortality (29.9% vs. 42.9%, P = 0.042), lower cardiopulmonary resuscitation (CPR) (10.2% vs. 21.4%, P = 0.013), lower continuous renal replacement therapy (CRRT) (24.2% vs. 40.0%, P = 0.009), and more advanced care planning decisions before death (87.3% vs. 66.7%, P = 0.013) than the former patients. The better ICU mortality (hazard ratio, 1.641; P = 0.035), lower CPR (odds ratio [OR], 2.891; P = 0.009), lower CRRT (OR, 2.602; P = 0.005), and more advanced care planning decisions before death (OR, 4.978; P = 0.007) were also associated with intensivist direct management in the multivariate cox and logistic regression analysis. Conclusion Intensivist direct management might be associated with better ICU outcomes than resident management under the supervision of an intensivist. Further large-scale prospective randomized trials are required to draw a definitive conclusion.