Drug use has been gradually declining among early and mid-adolescents since 1997. The timing of the decline roughly matches the introduction and diffusion (spread) of evidence-based prevention to middle schools in the United States. Evidence-based prevention refers to prevention programs, strategies, and policies that have been rigorously tested under research conditions and found to be effective in changing adolescent drug use behavior and attitudes. Federal and state funding for prevention is increasingly tied to community and school commitments to use only evidence-based strategies. However, local interpretation of what constitutes evidence-based prevention is highly variable, and subject to lack of knowledge about characteristics of evidence-based prevention, local politics and preferences for the status quo, and attempts to change parameters of evidence-based programs to fit perceived local needs. The current article reviews and synthesizes the characteristics and impact of evidence-based prevention, including the use of theoretical, process, and structural models; content guided by risk and protective factors for drug use; and settings and components of prevention delivery. Issues of adoption, implementation, and diffusion of evidence-based prevention are discussed. Gaps in knowledge of how to move the U.S. from research to practice are presented in terms of future researchable questions. Finally, the changing nature of prevention research is discussed as "action research," involving the negotiated partnership of the researcher and the community, and feedback of research results to communities to use as a planning tool for prevention.