We evaluated the intensity of antibiotic therapy in patients in whom the etiology of community-acquired pneumonia (CAP) was determined using newly available diagnostic techniques. For 1 year, we studied all patients admitted for findings consistent with CAP. Sputum and blood cultures, urinary pneumococcal and Legionella antigens, and viral polymerase chain reaction (PCR) were studied prospectively. Patients were stratified based on the final diagnoses: proven bacterial, presumptive bacterial, viral, fungal, undetermined, and uninfected. We determined the number of antibiotics given, duration of antibiotic therapy, and intensity of antibiotic use determined by antibiotic-days defined as the sum, in each patient, of all antibiotics given for CAP and the number of days given. Median duration and intensity of antibiotics were 12 and 18 days for proven, and 13 and 16.5 days for presumed bacterial CAP (P > 0.9). When positive viral PCR results were not disclosed to primary care physicians, antibiotic use was similar to that in bacterial CAP. However, in 11 cases, when positive viral PCR results were disclosed, duration and intensity of antibiotic use were reduced to 7 and 9 days, respectively (P = 0.05 and 0.08, respectively). Antibiotic use was similar in patients with bacterial pneumonia and those judged on clinical grounds to have likely nonbacterial infection. Despite obvious differences in clinical syndromes and final diagnoses, the intensity of antibiotic therapy was similar in all groups of patients admitted for CAP with the exception of those who were uninfected and whose primary care physicians were informed of a positive viral PCR.