摘要：Researchers, funders, and managers of health programs and interventions have become concerned about their long-term sustainability. However, most research about sustainability has not considered the nature of the program to be sustained. Health-related interventions may differ in their likelihood of sustainability and in the factors likely to influence continuation. I suggest a framework for analyzing the sustainability of 6 types of interventions: (1) those implemented by individual providers; (2) programs requiring coordination among multiple staff; (3) new policies, procedures, or technologies; (4) capacity or infrastructure building; (5) community partnerships or collaborations; and (6) broad-scale system change. Hypotheses for future research and strategies that program managers might use to achieve sustainability also differ by program or intervention type. Researchers and funders have expressed great interest recently in learning more about program sustainability. Do programs and other interventions that receive short-term funding continue after that investment ends? Can researchers show evidence for strategies that enhance the sustainability of health and human service programs? Questions about sustainability often arise for programs initially funded from an external government or foundation grant or contract. Since 2007, the National Institutes of Health (NIH) has fostered these interests by sponsoring annual conferences about research on dissemination and implementation, which include sessions on studying sustainability, especially of interventions developed and tested with NIH funding. 1 No agreed-upon definition exists for the term “health program sustainability.” Most sources agree that it refers to the continuation of activities or benefits for target recipients after an initial period of funding ends or following the initial implementation of a new program or procedure. According to one definition, “Sustainability is the continued use of program components and activities for the continued achievement of desirable program and population outcomes.” 2 (p2060) Terms such as “institutionalization” or “routinization” refer to embedding a new set of activities into the normal operations of an organization. Other authors have used the terms “continuation,” “maintenance,” and “durability” or their opposites, such as “discontinuation.” Sometimes sustainability is included as a subconcept under the broader concept of implementation, 3 but that framework does not emphasize the longitudinal perspective needed for sustainability research. Furthermore, factors that influence successful initial implementation are not necessarily the same factors that enable continued implementation into longer-term sustainability. A more detailed discussion of concepts affecting research on sustainability, as illustrated by 6 different definitions of outcomes or dependent variables for sustainability research, is available elsewhere. 2 Here I focus on sustainability of programs or interventions that have already been implemented in the target locations. A program is defined as a specified set of activities intended to address one or more identified outcomes, especially in the human services arena. 4 The focus of the sustainability study may also be termed an intervention, especially when it involves intervention into some aspect of the lives of recipients, to address their needs and enhance opportunities for productive lives and healthy behaviors. These programs and interventions take place in both clinical and community settings, although some are more likely to occur in a clinical health care setting, such as interventions delivered by individual providers; other programs more often occur in a community setting, such as community partnerships or broad-scale system change. In contrast to my usage here, sometimes the term “program” refers simply to a funding stream, such as a line item designating a program in the federal budget that allocates money for a broad purpose. Those programs may encompass many different types of activities in different locations, rather than a specified set of activities or components. Use of “program” to signify a diverse set of local activities makes it more difficult to assess local-level sustainability when no well-defined entity serves as the focus for research. In addition, when broad-scale systemic change is intended, the intervention strategy might include multiple programs focused on diverse forces interacting in an environment. 5–7 Some researchers define sustainability as the processes that lead to continuation of program activities after initial funding ends. 8,9 They appropriately suggest that later continuation of program components is influenced by processes that take place during program adoption decision-making and initial implementation. However, defining sustainability as this set of longer-term processes is not conducive to cumulative research measuring both sustained outcomes and the influences that lead to those outcomes. Several streams of research and research methodology have expanded the topic of sustainability in recent years: Systematic reviews of empirical research results about sustainability 10–13 ; Development of frameworks and conceptual models about factors likely to affect sustainability 10,13–16 (e.g., Schell et al., Center for Tobacco Policy Research, engaged experts in conceptual mapping to identify 8 factors believed to foster sustainability capacity 17 ); Discussion of methods needed to strengthen empirically based research about sustainability 2,13,18 ; Focus on sustainability as a set of processes likely to enhance the durability of programs and projects 8,9 ; Advice to funders and practitioners about the importance of planning early for sustainability and suggested strategies to enhance its likelihood 15,19 ; Development of related concepts, especially institutionalization and routinization 11,20–23 ; Numerous studies about the sustainability of specific programs or projects (reviewed by Scheirer 11 and by Stirman et al.13); Discussions of why sustainability is important and the need for more research about sustainability. 12,18,24 These streams overlap within and across specific articles; many sources contribute to several of these topics. To date, writers about health program sustainability have not focused on the nature of the program, project, or intervention to be sustained. Precisely what is being evaluated in studies of the extent of and processes underlying sustainability? Are factors influencing sustainability likely to differ for different types of interventions? For research about sustainability to contribute evidence-based propositions about how to sustain effective new programs, should the research differentiate among different types of interventions? Perhaps one size does not fit all for this research topic. To date, sustainability research and evaluation have treated the focus intervention as a black box whose contents are not expected to affect the research conclusions or the strategies needed to enhance sustainability. A similar weakness hampered evaluative studies of program impacts in the early days of evaluation research. 25,26 The subsequent widespread use of theory of change or logic models, along with measures of program implementation, now provide details about expected program components in most efficacy and effectiveness studies. I propose that sustainability research and advice to program managers should differ by type of program or intervention. The framework describes 6 types of interventions or programs distinguished by their structure: what they are and how they are implemented. They can be expected to differ in the processes and resources required to sustain them after initial funding ends. Hypotheses differ about key factors likely to foster the sustainability of each type of program or intervention; similarly, advice for program managers who seek to sustain their implemented programs should be tailored to the type of program. Advances in knowledge and practice about sustainability are more likely if researchers carefully define the nature and requirements of the entity to be sustained and then assess hypotheses appropriate to that type of intervention. The terms “program,” “project,” “intervention,” and “strategy” are often used interchangeably to refer to the focus of the sustainability research or practical advice. In some initiatives, a typical project may include several types of interventions, such as a community coalition that is both working toward school food policy changes and advocating for interventions by primary care providers to reduce childhood obesity. For these multicomponent projects, the likelihood of sustainability, and the factors affecting it, may differ considerably among those multiple components of a single project or site. Therefore, researchers and practitioners should plan appropriate data collection approaches and sustainability strategies for the diverse components of their projects. Legitimate questions may arise about the origin of this typology. It is not directly derived from previous theory, but rather developed from my reflections about recent analyses and advice to practitioners about health program sustainability. For example, the very helpful Program Sustainability Assessment Tool developed by Luke and colleagues contains rating scales for 8 factors believed to influence sustainability capacity, derived from the literature and a concept-mapping exercise. 14 (I served on an advisory committee aiding its development.) Yet no attention is given in that tool or its explanatory material to the possibility that these factors may relate differentially to different types of programs. For example, is the factor labeled political support equally important for individual physicians to continue prescribing a new medication and for the continuing work of community coalitions? A major new review of sustainability studies distinguishes among medical, mental health, and public health interventions but does not further characterize the structures of the interventions in the studies reviewed. 13 Previous theories and conceptual work about research translation processes have not developed specific typologies about program sustainability. For example, the new synthesis book, Dissemination and Implementation Research in Health: Translating Science to Practice , 27 provides no conceptual framework of intervention types, even for furthering research about initial implementation, which might have different predictors than does sustainability. Existing typologies focus most often on how specific program components might bring about behavioral change in target recipients, rather than illuminating diverse ways of sustaining interventions. 28–30 The frequently used diffusion of innovations framework includes characteristics of the intervention, such as complexity, relative advantage, and adaptability, but these are characteristics as perceived by specific stakeholders, not conceptualized as inherent structures of the intervention being considered. 31 The intervention typology has roots in congruence and open-system theories of organizational behavior, which emphasize the importance of fit among organizational and environmental components when assessing organizational change, 32–34 but no theory has applied these frameworks to sustaining health-related interventions. This typology is also related to the 6 types of sustainability outcomes or dependent variables suggested recently, 2 but does not imply a one-to-one correspondence between intervention type and intended outcomes. Space does not permit discussion of the full range of processes in a comprehensive approach to the dissemination and implementation of evidence-based practices to a large number of sites or individuals. 3,27,35 The broader topic of research-to-practice translation, or spread, of interventions requires attention to at least the following subcomponent processes after an intervention has been carefully developed and tested for efficacy: (1) dissemination of information about the intervention to potential adopters, who may be individuals or organizations working in either a clinical or a community context (or diffusion occurring by other methods, without an identified dissemination or marketing effort); (2) decision processes by potential adopters to start using the program or intervention; (3) implementation processes for adopters and adopter organizations to learn to use the intervention and to make other complementary changes, for example, in organizational processes needed to implement or adapt it appropriately; and (4) sustainability, the continued use of that intervention after an initial implementation period. Here I focus on the fourth process: how to characterize, research, and enhance the sustainability of beneficial interventions after initial implementation or short-term funding ends. Sustainability is likely to be influenced by all the earlier processes, but in different ways for different types of interventions or programs. I also do not consider fully here the environmental setting or context of the focus program, which is likely to have major influences on all these processes. For example, Kegler et al. describe several contextual factors that influenced the planning and implementation phases of community-based health promotion coalitions in California (they did not study sustainability), including history of collaboration, community demographics, economic conditions, geography, and community values and norms. 36 Other researchers found that support for implementing evidence-based practices differed between private and public mental health agencies. 37 Different intervention types often occur in different organizational contexts. For example, interventions implemented by individual providers might be found primarily in clinical settings, although these settings could range in scope from single-provider primary physician offices to much larger health care organizations. Future analyses might usefully identify and characterize different types of environments, to enable contingency research models for sustainability that consider both type of intervention and the favorability of environmental influences. 38
