The ultimate goal of HIV prevention interventions is to reduce the spread of HIV; however, the effectiveness of these programs is seldom assessed directly. Although direct measurement of an intervention's impact via HIV seroincidence monitoring is usually unfeasible, mathematical models can be used to estimate the number of infections averted by the intervention. This article describes three model-based summary measures of sexually transmitted HIV risk and discusses their relevance to HIV program evaluation in general economic efficiency analyses in particular. The calculation of these measures is demonstrated with an illustrative application to previously published data from an HIV prevention intervention for gay men.