In Colombia, health workers obtained a nasopharyngeal wash from 103 infants aged less than 12 months hospitalized for acute lower respiratory infection (ALRI) at the General Hospital of Medellin during April 1994 to April 1995 so researchers could determine the frequency of ALRI caused by respiratory syncytial virus (RSV) in hospitalized children. Immunofluorescence detected RSV infection in 43 (41.7%) patients. The presence of the following signs and symptoms allowed a clinical diagnosis of a viral infection: rhinorrhea, prolonged expiration, expiratory wheezing, interstitial infiltrates, and hyperinflation on chest radiographs as well as negative tests for 3 or 4 acute phase reactants. The physicians initiated antibiotic therapy (for 1-3 days) in 12 cases (27.9%) based on acute phase reactant findings who actually had an RSV infection. When the physicians learned that the laboratory confirmed RSV infection, they stopped antibiotic therapy. Antibiotics were continued in 16 (37.2%) other RSV infected infants, all of whom were less than 2 months old, due to mixed pneumonia (viral and bacterial). 39.4% of RSV-infected children whose clinical findings strongly suggested RSV received no antibiotics. None of these children or other ALRI patients with a viral disease suffered complications. They required less hospitalization time--since no further diagnostic tests were needed--than ALRI patients with a bacterial infection. Admissions for both ALRI and RSV infection peaked during November to January. RSV incidence peaked in January (23.3%). The leading reasons for hospitalization were pneumonia and bronchiolitis. These findings show that RSV diagnosis is useful and it lessens the indiscriminate use of antibiotics.