Evaluation of pneumonia severity scoring systems in nursing and healthcare-associated pneumonia for predicting prognosis: A prospective, cohort study

Abstract The usefulness of existing pneumonia severity indices for predicting mortality in nursing and healthcare-associated pneumonia (NHCAP) is unclear. This study compared the usefulness of existing pneumonia severity indices for predicting mortality in NHCAP and community-acquired pneumonia (CAP). Consecutive hospitalized pneumonia patients including NHCAP and CAP patients were prospectively enrolled between October 2010 and November 2017. Admission pneumonia severity was assessed using CURB-65, Pneumonia Severity Index (PSI), A-DROP, Infectious Diseases Society of America (IDSA)/American Thoracic Society (ATS) severe pneumonia criteria, and I-ROAD. The primary outcome was 30-day mortality. The discriminatory ability of each severity index was evaluated by receiver operating characteristic curve analysis. Overall, 828 patients had NHCAP, and 1330 patients had CAP. Thirty-day mortality was 12.8% and 5.6% in NHCAP and CAP patients, respectively. The area under the curve of PSI (0.717, 95% confidence interval 0.673–0.761) was the highest among all pneumonia severity indices, with significant differences compared with CURB-65 (0.651, 95% confidence interval 0.598–0.705, P = 0.02) and IDSA/ATS severe pneumonia criteria (0.659, 95% confidence interval 0.612–0.707, P = 0.03). The predictive abilities for 30-day mortality of the pneumonia severity indices, excluding PSI and I-ROAD, were significantly inferior for NHCAP than for CAP. PSI may be the most useful pneumonia severity score for predicting mortality in NHCAP. However, the predictive ability for mortality of each pneumonia severity score was worse for NHCAP than for CAP; therefore, the prognostic factors in NHCAP need to be identified for better management of NHCAP patients.