Implications of different SOFA score variants for outcome prediction in emergency department patients with an infection: a multicenter cohort study.
Abstract
Background
The Sequential Organ Failure Assessment (SOFA) score estimates mortality risk and ICU admission in Emergency Department (ED) patients with infection, but not all parameters are routinely available outside the ICU. Several adaptations have been proposed, yet their impact on prognostic performance remains unclear.
Methods
We conducted a post-hoc analysis of three prospective ED cohorts from two Dutch hospitals, including adult emergency department patients with confirmed infection. Seven a-priori selected SOFA score variants were evaluated: Original SOFA, AW-SOFA (automatic wrapper R-package), SF-SOFA (SpO2/FiO2-imputed), MICE-SOFA (multiple imputation chained equations), ΔSOFA (delta-SOFA), mSOFA (modified-SOFA), and eSOFA (electronic-SOFA). The primary outcome was a composite of in-hospital mortality and/or ICU admission.
Results
A total of 2,036 patients were included (median age 67 [IQR 55-74] years, 57% male). Severe outcomes occurred in 348 patients (17%); 236 were admitted to the ICU (12%), and 178 died in-hospital (9%). Predictive performance for severe outcomes was similar across most variants.
Conclusion
Seven SOFA variants showed broadly similar prognostic performance for severe outcome across methods. Methodological choices in SOFA implementation primarily influenced feasibility and the severity profile of classified patients rather than overall discrimination, supporting transparent reporting and context-specific selection of SOFA methods in ED studies.