Correspondence between hospital admission and the pneumonia severity index (PSI), CURB-65 criteria and comparison of their predictive value in mortality and hospital stay.
Infez Med. 2013 Jun;21(2):103-10
Authors: Varshochi M, Kianmehr P, Naghavi-Behzad M, Bayat-Makoo Z
Pneumonia severity assessment systems, such as the pneumonia severity index (PSI) and CURB-65, were designed to guide physicians to admit the patients involved to appropriate wards of hospitals. This study evaluated concordance rate of decisions leading to patients' hospitalization in accordance with PSI and CURB-65 criteria and comparison of the two systems' P-values in evaluating mortality and the hospitalization period of the patients in question. A total of 134 hospitalized patients with community-acquired pneumonia (CAP) were evaluated. Patients were classified on the basis of risk factors implicated in the PSI and CURB-65 systems. Prognostic P-values and indication measures of hospitalization for the two systems were then compared. Eighty-seven males (64.9%) and 47 females (35.1%) with a mean age of 64.23±19.82 (15-103) years were enrolled in the study. Based on the results of both systems, hospitalization was indicated in 112 cases (83.6%) and there was total agreement between the two systems in 61 cases (45.5%). There was no significant association between hospitalization duration in the two systems. However, both systems significantly predicted mortality within the hospitalization period with rather equal accuracies. Patients expired more frequently in the group with indication of hospitalization based on the PSI classes. However, there was no significant difference in the mortality between the two groups with and without admission indication according to the CURB-65 system. A considerable portion of our hospitalizations met the related criteria of the PSI/CURB-65. The two evaluation systems have near equal sensitivity and specificity for predicting mortality among hospitalized patients with CAP when the PSI class IV-V and CURB-65 score ≥ 2.
PMID: 23774973 [PubMed - indexed for MEDLINE]