• Chih-Chia Hsieh, Chao-Yung Yang, Chung-Hsun Lee, Chih-Hsien Chi, and Ching-Chi Lee.
• Department of Emergency Medicine, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan; Department of Medicine, National Cheng Kung University Medical College, Tainan, Taiwan.
• Am J Emerg Med. 2020 Feb 1; 38 (2): 282-287.

BackgroundThe Mortality in Emergency Department Sepsis (MEDS) score can be used to stratify ED patients with suspected infections according to mortality risk. However, it has yet to be externally validated for patients having bloodstream infections.MethodsWe retrospectively computed clinical information for the MEDS score, Pitt bacteremia score (PBS), Charlson comorbidity index (CCI), and McCabe-Jackson comorbid classification (MJCC) for adults with community-onset bacteremia. The MEDS score was validated by the comparisons with the following scoring systems: the PBS, CCI, MJCC, PBS plus MJCC, and PBS plus CCI. We evaluated goodness-of-fit statistics and c-statistics as measures of model calibration and discrimination, respectively.ResultsOf 2328 adults, a good calibration for 28-day crude mortality was obtained only in the MEDS score and PBS plus MJCC; a higher c-statistic (0.870, P < 0.001) were achieved by the MEDS score, compared to the PBS, CCI MJCC, PBS plus MJCC, and PBS plus CCI. A high c-statistic was observed in two combinative scoring system: the PBS plus CCI (0.855, P < 0.001) and PBS plus MJCC (0.843, P < 0.001). According to the Kaplan-Meier curves, 28-day crude mortality significantly differed between patients with scores equal to or higher than selected cutoff values and those with scores lower than selected cutoff values: 10 in the MEDS score and 5 in the PBS plus MJCC, respectively.ConclusionThe MEDS score is an excellent predictor of short-term outcomes in patients with community-onset bacteremia because it provides estimates with higher calibration and discrimination than those of the other scoring systems.Copyright © 2019. Published by Elsevier Inc.

30 keywords...

hide…