The epidemiology of influenza in the hospital is frequently confounded by failure to separate community-acquired from nosocomial transmission. An 83-year-old woman was hospitalized one day after returning from Asia with complications resulting from acute influenza A (H3N2) infection; she was the first culture-confirmed case in the region during the 1987-1988 influenza season, and her illness antedated other influenza cases in the area by at least four weeks. The patient shed virus at least four days after admission and transmitted influenza to her primary physician; both had received trivalent influenza vaccine four weeks earlier. Surveillance data from the 28 health care providers (HCPs) in contact with the index case (mean age: 34.5 years; median time of contact: four hours, none receiving vaccine) revealed no evidence of transmission as detected by paired type-specific complement-fixation antibodies and throat culture (20 subjects) or acute serologies and culture (7 subjects). No febrile respiratory illnesses were detected among other patients on the same ward, although three were reported among HCPs. Thus, neither secondary spread of influenza from infected patient to hospital HCPs nor nosocomial transmission apparently took place, although transmission did occur to the primary physician.