In an ideal world, decision makers and practitioners would have access to evidence-based programs and strategies to improve health and reduce health problems with the highest preventable burden. Decision makers would also have tools to guide their selecting and adapting appropriate approaches for specific contexts. In reality, much work needs to be done to achieve this goal, and many pressing public health problems exist where evidence is not yet fully established. Moreover, public health agencies at all levels face substantial fiscal constraints and challenges to improving health across varying contexts. Therefore, the need grows for evidence-based practices that afford maximum efficiency and effectiveness. Many agencies and organizations in the United States have acknowledged such challenges, including the White House (1), the Government Accountability Office (2), the US Department of Health and Human Services (3), the Congressional Budget Office (4), Healthy People 2020 (5), the Community Preventive Services Task Force (6), the National Academy of Sciences (7), and the Trust for America’s Health (8). International attention on evidence-based practice has mirrored that in the United States, including acknowledgment by the United Nations (9) and the World Health Organization (10).

Systematic reviews involve a critical examination of studies addressing a particular issue, using an organized method for locating and evaluating evidence. Many systematic reviews typically favor what Chen (11) refers to as the “Campbellian validity typology” — the use of research designs and methods that maximize internal validity to assess program efficacy. Identifying evidence-based practices by using this paradigm alone will often ignore factors critical to the successful integration of social and public health programs in real world settings. One exception is The Guide to Community Preventive Services (The Community Guide) — a repository of 1) recommendations and findings made by the independent, nonfederal Community Preventive Services Task Force about the effectiveness of community-based programs, services, and policies; and 2) the systematic reviews on which these recommendations are based (www.thecommunityguide.org) (6). The Community Guide reviews include research-tested and practice-based studies that use a range of study designs and assess both internal and external validity (6,12).

There is a need to assess the evidence for, categorize, and encourage additional study of strategies without enough evidence to undergo review by The Community Guide but that hold potential for public health impact. Practitioners and decision makers need tools to guide them toward selecting and evaluating the best available practices when published best practices are unavailable. Although systematic reviews identify what many public health professionals consider best practices, the public health field lacks a consensus definition and commonly accepted criteria for the designation best practice.

Consequently, the Office for State, Tribal, Local and Territorial Support at the Centers for Disease Control and Prevention (CDC) convened the CDC Best Practices Workgroup to develop a working definition of best practices, along with criteria and processes for classifying diverse practices in relationship to best practice. Through these efforts, the workgroup also aimed to encourage further development of practices that show promise for improving public health outcomes. We describe the development, structure, and use of the workgroup’s conceptual framework for creating a set of best practices. Our purposes in presenting the framework are to promote dialogue among scientists and practitioners about a consistent taxonomy for classifying the evidence for public health practices and to help researchers, practitioners, and evaluators show how their work contributes to building the evidence base for particular practices.

The CDC Best Practices Workgroup (see list of members in “Acknowledgments”) consisted of 25 CDC staff members with varying backgrounds (eg, epidemiology, behavioral science, program design and management, evaluation, policy development) and topic expertise (eg, chronic disease, infectious disease, injury prevention, environmental health). The workgroup reviewed the literature to find models and frameworks for classifying evidence, including best practices. The following question guided the review: What is known about the scope and definitions of best practices? The review included a keyword search of peer-reviewed and gray literature; workgroup members also identified relevant literature. The scope of the review was intentionally broad and included concepts related to evidence-based programs and policies as well as practice-based evidence.

The workgroup defined practice as “field-based or research-tested actions intended to effect a positive change.” Practices could include interventions, programs, strategies, policies, procedures, processes, or activities encompassing the 10 essential public health services and related activities (www.cdc.gov/nphpsp/essentialservices.html). This broad definition of practice includes activities designed to improve specific outcomes (eg, morbidity, mortality, disability, quality of life) or to focus on increasing the effectiveness and efficiency of program operations and services.

The review identified articles describing CDC initiatives that develop and disseminate recommendations for public health practice, such as Diffusion of Effective Behavioral Interventions (DEBI) (http://www.effectiveinterventions.org/en/HighImpactPrevention/Interventions.aspx); documents from The Community Guide (6,12); and non-CDC articles about identifying and using evidence to improve practice (1–5,7–10,13). Terms and descriptions found during the literature search appear in Table 1 (14–19). The workgroup abstracted key concepts, including gaps in defining and evaluating best practices and classifying other practices with varying amounts of evidence. A consensus definition of “best practice” was not found, but common elements were. In particular, we found that “best practice” and related terms do not refer to a static assessment or activity; rather, they refer to where, on a continuum, a particular practice falls at a given time. The review identified multiple ways to characterize this continuum or hierarchy, along with considerable variability in the number of stages or levels and in the rigor of methods used for identifying best practices.

The workgroup created a conceptual framework for planning and improving evidence-based practices by adapting and extending several streams of existing work related to developing a continuum of evidence (6,12,13,20–22). To broaden the framework’s usability, the workgroup developed criteria, definitions, and examples for key terms and formulated a series of questions to apply in assessing and classifying practices. The work was iterative and included frequent comparisons of the framework, definitions, criteria, and assessment questions with how these aspects of public health were discussed in the literature and with the extensive experience of workgroup members. We also continued to refine the products after the workgroup completed its tenure.

As a result of the review, the workgroup defined the term “best practice” as “a practice supported by a rigorous process of peer review and evaluation indicating effectiveness in improving health outcomes, generally demonstrated through systematic reviews.” The workgroup produced a conceptual framework (Figure) consisting of 2 interrelated components: public health impact and quality of evidence.

The public health impact component of the framework consists of the following 5 elements: effectiveness, reach, feasibility, sustainability, and transferability (Box). These elements (represented on the Figure’s vertical axis) are derived in part from the reach, effectiveness, adoption, implementation, maintenance (RE-AIM) model for translational research (20); the integrative validity model (11); and the systematic screening and assessment method (21).

Some elements of the public health impact component are more challenging to assess than other elements, such as sustainability and transferability. To address this issue, the workgroup developed questions that users can ask to determine the extent to which the practices they are developing or evaluating address each element (Box). The questions are not comprehensive, and some may not be relevant for all practices; however, having answers to these questions can help facilitate consistent interpretation of impact.

The quality-of-evidence component refers to where a practice lies on an evidence-based practice continuum (Figure). These elements represent 4 levels of evidentiary quality — weak, moderate, strong, and rigorous — that are represented on the horizontal axis. For example, field-based summaries or evaluations of progress with plausible but unproven impact are labeled as weak, whereas assessments of multiple studies or evaluations conducted by using systematic review methods are classified as rigorous (a complete set of definitions and examples is in Table 2) (22).

At the intersection of impact and quality of evidence, a continuum of evidence-based practice emerges, depicted by the arrow at the center of the Figure. This continuum represents the ongoing application of knowledge about what is working to strengthen impact in a given context. Building on the work by Brennan and colleagues (23), a lexicon was created for the continuum consisting of 4 stages — emerging, promising, leading, and best. In this conceptual framework, emerging practices include practices assessed through field-based summaries or evaluations in progress that show some evidence of effectiveness and at least plausible evidence of reach, feasibility, sustainability, and transferability. Emerging practices are generally newer, with a plausible theoretical basis and preliminary evidence of impact. These practices require more implementation and further evaluation to determine whether their potential impact can be replicated over time and in other settings and populations. Promising practices include practices assessed through unpublished intervention evaluations that have not been peer reviewed and that demonstrate some evidence of effectiveness, reach, feasibility, sustainability, and transferability. Promising practices have been evaluated more thoroughly than emerging practices and may include practices with higher quality of evidence and lower impact or with lower quality of evidence and higher impact, where decisions related to application will likely depend on context. Leading practices include practices assessed through peer-reviewed studies or through nonsystematic review of published intervention evaluations that show growing evidence of effectiveness and some combination of evidence of reach, feasibility, sustainability, and transferability. Best practices adhere to the most rigorous assessments in the continuum, including systematic reviews of research and evaluation studies, which demonstrate evidence of effectiveness and growing evidence of reach, feasibility, sustainability, and transferability (eg, The Community Guide [12] and the HIV/AIDS Prevention Research Synthesis Project).

These classifications are hypothesized to be dynamic and can change over time for a given practice. For example, a practice may be assessed as promising at one point and be promoted to leading or best as stronger evidence is developed. Conversely, practices once considered “best” may become outdated as the field and environmental conditions evolve. The dynamic quality of these classifications highlights the importance of regularly re-evaluating practices at all points on the continuum, including updating systematic reviews of best practices. Such updates, incorporating all emerging evidence, are a critical component of the work of the Community Preventive Services Task Force and other review groups (6). By encouraging evaluation at each stage and supporting continued evaluation to move worthy practices up the evidence continuum, the framework aids quality and performance improvement (24).

The framework as it now appears has several limitations: 1) it is conceptual and has not been fully evaluated, and 2) the two axes will be challenging to measure and balance against each other. Classifying a particular practice reliably using the framework depends on the background, perspective, and skill of the individuals assessing the practice.

Because substantial variation exists in practices needed to address various public health challenges, the questions and guidance used to make the designation of emerging, promising, leading, or best require ongoing refinement. Additional tools are needed to guide users in consistently examining the evidence related to a given practice in diverse circumstances. Helping users to develop a clear definition that addresses scope and boundaries for each practice being assessed is important because the development and implementation of public health practices often vary from one setting or population to another. The framework would benefit from additional input by diverse users and stakeholders to inform an iterative process of improvement and refinement over time.

To address these limitations, tools were developed to apply the framework, to determine quality and impact, and to begin to address the complexity of combining 5 elements. The framework and tools are undergoing pilot testing in numerous CDC programs. Pilot testing will assess the validity, intrarater and interrater reliability, and utility of the framework. Results of this pilot testing and further application of the framework will be published.

CDC is congressionally mandated to support the Community Preventive Services Task Force in developing the systematic reviews and related evidence-based recommendations found in The Community Guide (6). CDC also supports development of other systematic review activities (eg, the DEBI project) to identify best practices through rigorous assessment, advance public health science, and promote translation of interventions supported by the highest levels of evidence. The conceptual framework presented here builds on and complements these efforts by offering a practical approach designed for use by public health practitioners, evaluators, and researchers. The framework offers no justification for implementing less than a best practice in cases where a best practice is known and appropriate to address a public health problem. However, in cases where no practices have achieved best practice status, the framework supports the use of the best available practice for a given health problem and context while 1) making sure users recognize where that practice sits on the evidence continuum and 2) encouraging evaluation of impact, including publication of findings to contribute to the overall evidence base for that practice. A key goal for this framework is to improve public health programs by building practice-based evidence (20).

The value of the framework lies in offering a common lexicon and processes to strengthen the evidence for public health practice. The Federal Interagency Workgroup for Healthy People 2020 has adapted the evidence continuum components of the framework as developed by the CDC workgroup. Other stakeholders may use the framework, definitions, and criteria to communicate more clearly about the continuum of evidence for a wide spectrum of programs, including those that do not currently qualify as best practices but may benefit from additional study. Initial feedback from scientists, evaluators, and practitioners, both internal and external to CDC, supports both the face validity of and the need for this framework.

Applying the elements identified in the conceptual framework might assist with program planning, evaluation, and dissemination. The framework could potentially increase transparency and accountability for public health programming by clarifying the components and definitions of evidence. It could help practitioners and evaluators identify gaps in the evidence for a given practice and determine how their work contributes to building the evidence base. Practitioners might attend to the framework’s effectiveness criteria as part of logic model development to help identify potential gaps between activities and intended outputs and outcomes. Several of these criteria already have been adopted in protocols to identify practice-based strategies for more rigorous evaluation (21). The framework’s inclusion of feasibility and transferability could help facilitate adoption by increasing the likelihood that a practice will translate to other settings.

The conceptual framework is offered to promote dialogue among researchers, evaluators, practitioners, funders, and other decision makers. With this framework, we seek to foster a shared understanding for assessing standards of evidence and motivate organizations to implement and improve strategies that move practices along the continuum from emerging to best.