Louise Holm

Objectives: To investigate the prognostic implication of the SOFA score on all acutely admitted medical patients, to see how well it could predict 30-day mortality and ICU- admission.

Main outcome measures: Discriminatory power was calculated as AUROC. Calibration was assessed using Hosmer-Lemeshow goodness-of-fit.

Results: AUROC for 30-day mortality was 0.68, (95% CI, 0.64 to 0.71) and ICU admission 0.71, (95% CI, 0.66 to 0.76). Goodness of fit for 30-day mortality and ICU admission was acceptable.

Conclusions: The SOFA score showed average ability to predict 30-day mortality and ICU admission with acceptable calibration. When substituting GCS with AVPU the performance of the SOFA score was unacceptable.