FIGURE 3-3 Five phases of prevention research.

SOURCE: Reprinted with permission, from Kellam and Langevin, 2003. Copyright 2003 by Prevention Science.

universal application that is independent of time, place, or context (Tang et al., 2003). Falsifiability refers to the possibility that an evaluation could determine, if relevant, that the intervention is ineffective (Tang et al., 2003).

An evidence-based medicine approach has been adopted by federal and scientific institutions to guide obesity treatment in adults (NHLBI, 1998). However, efforts to apply its principles to identifying effective interventions for other areas of disease prevention and health promotion have met with varying degrees of success (Osaka Declaration, 2001; McQueen, 2002; WHO, 2003; Victora et al., 2004). Indeed, it has been suggested that clinical decisions may have a relatively small impact on health outcomes compared to changes in the social environment, and that broadening evidence-based medicine beyond clinical policy decision-making—to public health decision-making—often has the potential to produce a larger beneficial impact on the health of populations (Heller and Page, 2002).

As the public health and health promotion disciplines have evolved, evidence-based public health has become the goal with a knowledge base that includes disease frequency and distribution; correlates, determinants and consequences of disease; and the safety, efficacy, effectiveness, and cost-effectiveness of a range of interventions (Victora et al., 2004). But given the complex environment in which multiple social, economic, cultural, and political elements interact to produce change in population-wide problems such as obesity, causality may not always be established for the