Study objectives: (1) To define clinical factors that could justify hospital admission among patients with community-acquired pneumonia (CAP) with risk classes of I or II. (2) To determine the positive predictive value of the pneumonia severity index as the sole indicator for detecting inappropriate hospitalizations among patients with CAP.
Design: Retrospective observational study.
Setting: University of Louisville Hospital and the Veterans Affairs Medical Center of Louisville, KY.
Patients: Consecutive adult patients fulfilling the criteria for CAP who were admitted to the hospital between October 1997 and May 2000.
Measurements and results: The medical records of hospitalized patients with CAP having a risk class of I or II were identified and further reviewed to determine whether there existed a clinical basis to justify hospitalization. Of a total of 328 patients, 86 had a risk class of I or II. Among these, 72 had clinical factors that justified their hospital admission. These factors, in frequency of occurrence, included the following: medical conditions other than CAP that required hospitalization, 31 patients (43%); social needs, 13 patients (18%); oral intolerance, 10 patients (14%); failure of outpatient therapy, 10 patients (14%); noncompliance, 6 patients (9%); suspicion of sepsis, 1 patient (1%); and hypoxemia, 1 patient (1%). Among the 86 patients with low risk classes (ie, classes I or II), 14 were found to be inappropriately hospitalized, yielding a positive predictive value of 16%.
Conclusions: The pneumonia severity index, used as the sole indicator for detecting inappropriate hospitalizations, has an unacceptably low positive predictive value. This is due primarily to the severity of comorbid conditions requiring in-hospital care in patients with a nonsevere pneumonia. According to our study, the pneumonia severity index should not be used as the sole indicator with which to define inappropriate hospitalization.