To assess the cost-effectiveness of a cue to influenza vaccination provided by community pharmacists, a decision tree was constructed of the consequences of implementing a pharmacy-based vaccine-advocacy program, based on experience gained in an experiment involving three community pharmacies in Durham County, North Carolina. The model used morbidity and mortality assumptions derived from the infectious-disease literature and cost assumptions based on 1990-91 Medicare Part A and Part B reimbursement costs. This analysis suggests that if Medicare reimbursed pharmacists for advising 100,000 patients at risk to accept influenza vaccine through vaccine-advocacy messages, for an apparent expenditure of $110,000, the increased rate of influenza vaccinations would avert 139 hospitalizations and 63 deaths, and actually yield Medicare a net savings of $280,588. These calculations probably underestimate the benefit to society of a pharmacy-based vaccine-advocacy program, because only direct costs to the single government agency were computed and no cost was attributed to death or lost earnings.