Outbreaks of influenza A (H3N2, A/Shanghai/11/87-like) occurred in two partially (60% and 79%) vaccinated nursing home populations in January 1988. A retrospective cohort study using chart review was designed to assess the effectiveness of influenza vaccination and amantadine prophylaxis (100 mg per day) in controlling the outbreaks and to determine the amantadine susceptibility of influenza viruses isolated from case-patients. The point estimate of vaccine efficacy in preventing influenza-like illness was -33% (95% confidence interval -115% to 18%). However, 9% of vaccinated case-patients died within 14 days after onset of influenza-like illness compared with 26% of unvaccinated case-patients (relative risk = 0.4, 95% confidence interval 0.1-1.0). There was no significant difference in illness severity among case-patients who became ill before amantadine prophylaxis was started (n = 84) compared with those who became ill while taking amantadine (n = 34). Four virus isolates obtained before amantadine prophylaxis was started demonstrated 52-68% inhibition by 1 microgram/ml of amantadine; by comparison, six isolates (resistant viruses) obtained from residents who became ill while taking amantadine demonstrated 1-18% inhibition. The resistant viruses had four different RNA sequences in the gene coding for the M2 protein transmembrane region. Three resistant viruses with identical RNA sequences were isolated from residents living in contiguous rooms who had onset of signs and symptoms during a 6-day interval. Further studies are needed to determine how frequently and under what circumstances resistant viruses occur when antiviral agents are used to control institutional influenza A outbreaks. Strategies for antiviral agent administration that limit the emergence and transmission of resistant virus strains may be needed.