The Food and Drug Administration (FDA) must manage the delicate balance between speed and safety in the evaluation and approval of new drugs. To explore how these competing obligations might be formalized with respect to the AIDS epidemic, we present a simple decision-theoretic optimization model of the clinical trials process against a backdrop of HIV transmission. Our framework sheds light on such issues as the economic consequences of decisions, the potential savings of a new therapy, the costs to society of delay, the value of better information, and how and when to undertake a clinical trial. We believe that this article represents a first effort to unravel the tangled web of epidemiological, economic, and statistical considerations that plague this policy issue.