Predicting the need for hospitalization of ambulatory patients with pneumonia

To develop clinical guidelines to identify patients with pneumonia who might be safely treated as outpatients.

Retrospective chart review to derive guidelines, with subsequent prospective validation.

Initial review completed for patients seen in the emergency room (ER) of a university hospital and a community-based internal medicine practice. Validation conducted in the ERs of a university hospital and a community teaching hospital.

Individuals aged 16 years and older presenting with newly diagnosed pneumonia. Follow-up obtained through mail or telephone contact and chart review.

None.

During the initial retrospective review, 141 pneumonia patients without obvious reasons for hospital admission were identified and then classified as hospitalization necessary or unnecessary. Of these patients, 33 were classified as requiring admission. Multivariate analysis identified five variables that differentiated low-risk from high-risk study patients. These variables (and their relative weights) were: serious comorbid illness (3 points); preexisting lung disease (2 points); multilobar lung involvement by the disease by chest x-ray (2 points); observed or likely aspiration (2 points); and symptom duration of less than 7 or greater than 28 days (1 point). Risk scores were calculated using these variables. Patients with low scores (0-2 points) rarely had complications, and only one of 53 such patients (2%) was judged to require hospitalization. In contrast, patients with high scores (greater than or equal to 6 points) had frequent complications and 20 of 29 (69%) were felt to need hospitalization. Similar results were found during the validation phase.

Clinical findings appear to help distinguish patients who need admission for treatment of pneumonia from those who do not. If validated in other settings, the clinical utility of these guidelines in assisting decision making about hospitalization should be determined.