CURB-65 Pneumonia Severity Assessment Adapted for Electronic Decision Support.
Chest. 2010 Dec 16;
Authors: Jones BE, Jones J, Bewick T, Lim WS, Aronsky D, Brown S, Boersma WG, Van der Eerden M, Dean N
ABSTRACT BACKGROUND: Accurate severity assessment is crucial to the initial management of community-acquired pneumonia. The CURB-65 score contains data routinely entered in electronic medical records and thus electronically calculable. Aim: To determine whether an electronically generated severity estimate using CURB-65 elements as continuous and weighted variables better predicts 30-day mortality than the traditional CURB-65. METHODS: In a retrospective cohort study at a United States university-affiliated community teaching hospital, we identified 2,069 patients 18 years or older with community-acquired pneumonia confirmed by radiographic findings in the emergency department. CURB-65 elements were extracted from the electronic medical record and 30-day mortality was identified with the Utah Population Database. Performance of a severity prediction model using continuous and weighted CURB-65 variables was compared to the traditional CURB-65 in the US derivation population and validated in the original 1,048 patients from the CURB-65 international derivation study (Lim 2003). RESULTS: The traditional, binary CURB-65 score predicted mortality in the US cohort with an area under the curve (AUC) of 0.82. Our severity prediction model generated from continuous, weighted CURB-65 elements was superior to the traditional CURB-65, with an out-of-bag AUC of 0.86 (p<0.001). This finding was validated in the international database, with an AUC of 0.85 for the electronic model compared to 0.80 for the traditional CURB-65 (p = 0.01). CONCLUSIONS: Utilizing CURB-65 elements as continuous and weighted data improved prediction of 30-day mortality, and could be used as a real-time, electronic decision support tool, or to adjust outcomes by severity when comparing processes of care.
PMID: 21163875 [PubMed - as supplied by publisher]