Abstract
Background: Accurate severity assessment is crucial to the initial management of community-acquired pneumonia (CAP). The CURB-65 (confusion, uremia, respiratory rate, BP, age ≥ 65 years) score contains data that are entered routinely in electronic medical records and are, thus, electronically calculable. The aim of this study was 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 US university-affiliated community teaching hospital, we identified 2,069 patients aged 18 years or older with CAP confirmed by radiographic findings in the ED. 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 with the traditional CURB-65 in the US derivation population and validated in the original 1,048 patients from the CURB-65 international derivation study. 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<.001). This finding was validated in the international database, with an AUC of 0.85 for the electronic model compared with 0.80 for the traditional CURB-65 (P =.01). Conclusions: Using 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.
| Original language | English |
|---|---|
| Pages (from-to) | 156-163 |
| Number of pages | 8 |
| Journal | Chest |
| Volume | 140 |
| Issue number | 1 |
| DOIs | |
| Publication status | Published - 1 Jul 2011 |
| Externally published | Yes |
Bibliographical note
Funding Information:Funding/Support: This research is supported by the Deseret Foundation of Salt Lake City, UT.
Funding Information:
Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Lim has received an unrestricted educational grant from Wyeth/Pzier for clinical research. Dr Dean has received an unrestricted educational grant from Pfizer. He has also served on advisory boards for Forest, Merck, Ortho-McNeil, and Advanced Life Sciences. The remaining authors have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.