Abstract
This study assessed the performance of Sequential Organ Failure Assessment, Logistic Organ Dysfunction Score and Multiple Organ Dysfunction Score in outcome prediction in severe sepsis. A total of 528 consecutive patients with a diagnosis of severe sepsis were enrolled from two surgical intensive care units of university hospitals in China. Clinical and laboratory data of patients were collected and admission and maximum values of each scoring system were calculated. Areas under the receiver operating characteristic curve, which were used to assess discrimination, were 0.80, 0.83 and 0.74 for admission Sequential Organ Failure Assessment, Logistic Organ Dysfunction Score and Multiple Organ Dysfunction Score respectively, and 0.91, 0.93 and 0.86 for corresponding maximum values respectively. Calibration assessed by the Hosmer-Lemeshow statistic was better with admission (χ2=18.2) and maximum Logistic Organ Dysfunction Score (χ2=19.6) than with admission (χ2=98.1) and maximum Multiple Organ Dysfunction Score (χ2=30.9). Brier Scores, indicating the overall performance of the scores, were 0.18, 0.17 and 0.22 for admission Sequential Organ Failure Assessment, Logistic Organ Dysfunction Score and Multiple Organ Dysfunction Score respectively, and 0.12, 0.10 and 0.15 for their maximum counterparts respectively. This study found good performance of both admission Sequential Organ Failure Assessment and Logistic Organ Dysfunction Score in severe sepsis, and a slightly weaker performance of admission Multiple Organ Dysfunction Score. Since poor calibration was observed in Logistic Organ Dysfunction Score and Multiple Organ Dysfunction Score, we suggest further study of customisation of these scores in critical illness with severe sepsis.