The U.S. government continues to improve its plans for protecting civilians and soldiers from attacks with biological weapons. Part of this effort focuses on developing strategies that recognize the difficult choices to be made in using and deploying resources. This paper presents a risk- and decision-based framework-derived from the field of Bayesian statistics-for developing strategies that facilitate managing the risks of biological agents. The framework recognizes the significantly different attributes of potential biological weapons and offers a strategy for improving communication to effectively coordinate national biopreparedness efforts. The framework identifies generic decisions related to routine immunization, response planning, stockpiling vaccines and therapeutic agents, surveillance choices, containment, emergency response training, research, media and communications preparations, information management, and policy development. This paper provides a straw man to be used in wargames, exercises, practices, etc., at all levels of government. Given the attention on anthrax following the 2001 attacks, this paper applies the framework to managing the risks of anthrax to provide an illustrative example. The example demonstrates that by organizing information at this level, decision makers can quickly understand the critical connections between different options (e.g., vaccinating with a new vaccine requires an investment in research; research might increase the opportunities for breaches of containment). With respect to managing the risks of an attack with anthrax, this analysis suggests the need for creation of a comprehensive national management plan that includes quantitative evaluation of resource investments. The authors conclude that the government should adopt a process-based on decision science and using the power of decision trees as an analytical tool-to develop a strategy for managing the risks of bioterrorism. Using this type of approach, the government can better characterize the costs, risks and benefits of different policy options and ensure the integration of policy development. Additionally, confirmed use and refinement of decision trees during exercises will provide analysis of the long-term consequences of decisions made during an event and give policymakers insights to improve initial decisions.