Weekends affect mortality risk and chance of discharge in critically ill patients: A retrospective study in the Austrian registry for intensive care

Abstract

Background: In this study, we primarily investigated whether ICU admission or ICU stay at weekends (Saturday and Sunday) is associated with a different risk of ICU mortality or chance of ICU discharge than ICU admission or ICU stay on weekdays (Monday to Friday). Secondarily, we analysed whether weekend ICU admission or ICU stay influences risk of hospital mortality or chance of hospital discharge. Methods: A retrospective study was performed for all adult patients admitted to 119 ICUs participating in the benchmarking project of the Austrian Centre for Documentation and Quality Assurance in Intensive Care (ASDI) between 2012 and 2015. Readmissions to the ICU during the same hospital stay were excluded. Results: In a multivariable competing risk analysis, a strong weekend effect was observed. Patients admitted to ICUs on Saturday or Sunday had a higher mortality risk after adjustment for severity of illness by Simplified Acute Physiology Score (SAPS) 3, year, month of the year, type of admission, ICU, and weekday of death or discharge. Hazard ratios (95% confidence interval) for death in the ICU following admission on a Saturday or Sunday compared with Wednesday were 1.15 (1.08-1.23) and 1.11 (1.03-1.18), respectively. Lower hazard ratios were observed for dying on a Saturday (0.93 (0.87-1.00)) or Sunday (0.85 (0.80-0.91)) compared with Wednesday. This is probably related to the reduced chance of being discharged from the ICU at the weekend (0.63 (0.62-064) for Saturday and 0.56 (0.55-0.57) for Sunday). Similar results were found for hospital mortality and hospital discharge following ICU admission. Conclusions: Patients admitted to ICUs at weekends are at increased risk of death in both the ICU and the hospital even after rigorous adjustment for severity of illness. Conversely, death in the ICU and discharge from the ICU are significantly less likely at weekends.