In this analysis, we assessed the costs and benefits of postattack and preattack strategies for an anthrax release of a greater magnitude than that previously experienced in the United States. The main finding of our analyses of postattack strategies is that use of vaccine plus antibiotic prophylaxis is the most effective and least expensive therapy. The savings associated with preventing cases of inhalational anthrax offset the cost of using both vaccination and antibiotics. This finding is robust even with reasonable changes in our estimates of the cost and efficacy of the vaccine and antibiotics. Our most important finding about preattack vaccination is that the net health benefit and cost-effectiveness depended critically on the probability of an attack and on the proportion of the population exposed during the attack. For a large metropolitan U.S. city, vaccination provides reasonable value for the health care dollar only when the probability of clinically significant exposure reaches about 1 in 200 (for example, when the probability of attack is 0.01 and the probability of exposure during an attack is 0.5, the joint probability of clinically significant exposures would be 0.005 or 1 in 200). Our findings highlight the inherent difficulties in decision making about anthrax vaccination. Several factors influence the probability that an individual will receive a clinically significant exposure during an attack, including the quantity of spores released, method of dissemination, and environmental factors (such as geography, wind conditions, and time of day of the dispersal) (4 - 6,23,65). Although it is difficult to judge the likelihood of a release and the probability of exposure given a release, clearly some individuals are at higher risk than others. Our finding that vaccination provides a net health benefit at even relatively low probabilities (1 in 500) may help decision-makers assess the desirability of vaccination of military and emergency services personnel, who are probably at greater risk for exposure than the general population. If a vaccine with fewer adverse reactions became available, the probabilities of exposure at which there may be net benefit would be lower.