The L.E.A.D. framework adapts an EBPH typology recommended by the International Obesity Task Force (3,9) for specifying questions (Figure 2). The “why” questions help decision makers frame reasons for considering or taking an action based on issues in their specific locality, region, or situation, which may include posing questions to assess baseline status or resources of the relevant population or setting. The “what” questions focus on selection of specific programmatic or policy initiatives and may include assessments of the potential effectiveness or value of approaches designed for specific settings (eg, schools, worksites, faith organizations) or subpopulations (eg, children of different ages, ethnic minority populations, low-income populations). “How” questions prompt for information about implementation issues, including resources required, how effects can be sustained over time, and factors that determine the generalizability or transferability of an approach tested in one setting to another setting.

The L.E.A.D. framework identifies several different types of evidence and study designs that can be useful for informing decisions about obesity prevention and provides resources and explanations in the report narrative and appendices. L.E.A.D. does not imply lowering standards for the quality of evidence used in public health compared with medicine. Rather, it introduces the concept that broader and different standards are needed to account for the complexity and practicalities associated with issues that arise in obesity prevention and other public health problems (eg, tobacco use, environmental health issues). Evidence sources may be quantitative or qualitative or a combination of both and may come from academic research, program evaluations, surveys, polls, reports, or policy documents related to obesity or to other public health issues from which parallels can be drawn. Specific study designs and methods identified in the report include experiments and quasi-experiments, qualitative research, mixed methods, evidence synthesis methods, parallel evidence, and expert knowledge. The point is to be systematic but much more inclusive when determining what constitutes useful evidence related to a particular question. Evidence quality is then to be evaluated by standards appropriate to that type of evidence rather than by a single standard. Examples of existing criteria for assessing quality of evidence from these different methods are provided.

One way that EBPH differs from EBM is that it relies less on randomized controlled trials and more on approaches that assess external validity. Study design cannot be the sole criterion for whether evidence is useful. A randomized controlled trial is the most rigorous design for hypothesis testing (10) but is not feasible for many examples of obesity interventions because the evaluator cannot randomly assign exposure (eg, a policy). Randomized designs also may provide incomplete information if the experiments evaluate artificial scenarios that have limited or only partial relevance to what happens in reality. Partial relevance might occur if the trial manipulates only 1 or 2 of several variables that interact in a dynamic manner to affect an outcome. In general, studies of obesity prevention have tended to overemphasize internal validity (eg, well-controlled efficacy trials) while giving sparse attention to external validity (eg, the translation of science to the various circumstances of practice) (11,12). The work of Klesges et al shows that some contextual variables (eg, cost, program sustainability) are missing entirely in the peer-reviewed literature on obesity prevention (13). Conversely, rigorous evaluations of nonrandomized, “natural experiments” can be informative for many obesity prevention questions. Natural experiments refer to naturally occurring circumstances in which different populations are exposed or not exposed to a potentially causal factor (eg, a stringent new school food policy) such that the situation resembles a true experiment in which study participants are assigned to exposed and unexposed groups (14). These types of studies often involve “messier” study designs (ie, complex, multilevel, multisector interventions) and suggest the need to take a broader perspective in identifying evidence and pay greater attention to external validity and situational or population-specific variables. The real world is actually “messy” or, more formally, complex. In a natural experiment, the strongest design possible (internal validity) is essential, and elements of external validity must be addressed. The L.E.A.D. framework guidance is designed to help incorporate this complexity into evidence rather than controlling for it, which detracts from and may completely remove contextual relevance.

The ultimate goal of the L.E.A.D. framework process is to assemble evidence in a way that is useful to decision makers. The framework recommends a standard template that can be used to report results to decision makers, which prompts for 1) a statement of the question, 2) a transparent description of the strategy used to locate the evidence, 3) a table reporting the evidence, and 4) a summary of the evidence organized as answers to the EBPH-derived questions (Figure 2). Because policy questions often focus on selecting the most feasible intervention, especially detailed guidance is provided about how to interpret and assemble evidence related to “what” questions. This guidance includes a discussion of how one might apply theory or program logic and a systems lens in interpreting evidence, considerations for weighting different types of evidence, and potential ways to blend information from disparate sources and to evaluate effects. Potentially useful tools and frameworks for grading and assembling different types of evidence are identified and include tools used in EBM: meta-analytic approaches to determine intervention effect size and the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) system for evaluating factors affecting the strength of recommendations. Other tools identified of particular relevance to public health applications include realist reviews that use mixed methods to assess intervention effectiveness, the systematic review approach of the Guide to Community Preventive Services framework for translating evidence into recommendations, the RE-AIM framework (Reach, Effectiveness, Adoption, Implementation, and Maintenance) for translating research into practice, the Health Canada risk assessment and management framework, the International Obesity Task Force obesity prevention portfolio approach for selecting a set of interventions, and the Green and Kreuter framework for identifying program components and interventions (“matching, mapping, pooling, and patching”) (9,15-19).

In addition to the systems perspective, recognizing opportunities to generate new evidence is also recommended as a theme of the L.E.A.D. framework. Such opportunities might arise at any stage of the process. Generation of new evidence is critical not only because of the dearth of suitable evidence about obesity-relevant environmental and policy changes but also to keep the evidence base current with the dynamics of the obesity problem. The “why,” “what,” and “how” questions in Figure 2 should guide the type of evidence generated. EBPH approaches to filling evidence gaps include program evaluations and natural experiments. This is termed “practice-based evidence” and may also include pre-evaluations or “evaluability assessments” of promising programs and continuous quality improvement. The case study (Box) of data collection in support of school-based physical activity interventions by evaluating programs in northern California communities illustrates the L.E.A.D. concept of using practice settings to increase the evidence base.

The Bridging the Evidence Gap report directs recommendations to decision makers in the policy and programmatic arenas as well as to those who fund, generate, and publish evidence about obesity prevention and other complex public health challenges. Central themes are to apply the L.E.A.D. framework as a guide in the use and generation of evidence and to incorporate systems-thinking into research activities. The report also recommends the development of resources to support evidence-based public policy decision making and research, including researcher training, compendiums of knowledge, registries of implementation experience, and guidance on standards for evidence evaluation where they are lacking. The need for a public-private consortium to take up dissemination, support for, and further development of the L.E.A.D. framework is emphasized.

In the approximately 2 years since the Bridging the Evidence Gap report was released, it has gained visibility among potential users. The report page on the IOM website has generated more than 17,000 page views. The L.E.A.D. framework has been presented at national and international meetings and cited in journal articles and policy documents. Scanning approximately 40 identified citations indicates that most have involved referencing the report in support of the importance of developing comprehensive multistakeholder and multisectoral strategies, taking a systems perspective, or using expanded approaches to evidence-gathering or choice of study designs. Applications of the framework reflected in published documents include a CDC Division of Nutrition, Physical Activity, and Obesity fact sheet that explains how the framework relates to potential uses of their research and practice-based initiatives, evidence sources, and guidance documents (30); adaptation of L.E.A.D. concepts to describe implications of a systems approach for policy and actions to address the global obesity epidemic (31); use of L.E.A.D. concepts to justify and propose a design for a large-scale demonstration and evaluation of a comprehensive community-based obesity prevention strategy beginning in early life (32); use of L.E.A.D. elements as the primary method for a review of progress made by the food industry, governments, and schools in implementing recommendations of 2 earlier IOM reports on childhood obesity (33,34); and extensive use of L.E.A.D. framework perspectives and evidence review guidance in the IOM report that recommends a set of systems-oriented and interrelated strategies to accelerate progress in obesity prevention (1).

What does this mean for obesity prevention and for advancing appreciation for the science and practice of EBPH? The answer depends on further use of frameworks such as L.E.A.D. In an ideal sense, L.E.A.D. could become a transformative and integrative EBPH paradigm and tool, as intended by the IOM study committee that developed it. The transformational aspect is the positioning of evidence needs in a public health context and demonstrating that rigor and relevance can be achieved using EBPH concepts and tools. Advances in obesity prevention will depend in part on articulating the value of multiple and varied types of information for answering policy and practice questions. If widely adopted and used, L.E.A.D. could become a critical component of identifying and using evidence-informed strategies for achieving national health objectives (35). Such use is likely to better link the practice of public health with the science of public health.