Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
11 Implementation and Dissemination of Prevention Programs P art II illustrates the substantial progress made in prevention science since 1994. It describes numerous efficacious or effective preven- tion programs (Chapters 6 and 7), as well as the cost-effectiveness of many of these programs (Chapter 9). It also demonstrates numerous methodological advances that increase confidence in the reliability of evi- dence that provides a strong basis for believing that the mental, emotional, and behavioral health of the nationâs young people could be significantly improved if evidence-based programs and policies were widely used (Chapter 10). Thus far, however, preventive interventions have generally not been widely implemented in schools and communities (Ennet, Ringwalt, et al., 2003; Gottfredson and Gottfredson, 2002; Hallfors and Godette, 2002; Wandersman and Florin, 2003) and have done little to reduce behavioral health problems in American communities (Chinman, Hannah, et al., 2005; Sandler, Ostrom, et al., 2005). While sustained, high-quality implementation by communities is essen- tial to achieving greater public health impact from the available tested and effective preventive interventions (Elliott and Mihalic, 2004; Glasgow, Klesges, et al., 2004; Spoth and Greenberg, 2005), implementation of exist- ing programs alone is unlikely to be sufficient. Implementation must also include development and evaluation of research-based adaptations of pro- grams to new cultural, linguistic, and socioeconomic groups; evaluation of approaches that have broad community endorsement; and implementation of policies and principles that support healthy development. This chapter begins with a discussion of alternative implementation approaches. It goes on to review examples of experience with implementa- 297
298 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS tion of existing prevention programs, as well as a number of challenges to implementation. The chapter then describes strategies that can complement the implementation of evidence-based interventions. Next is a discussion of research needed to increase understanding of and support successful imple- mentation. The final section presents conclusions and recommendations for moving implementation forward. IMPLEMENTATION APPROACHES A major implementation issue is the balance between delivering an evidence-based program as developed and adapting a program to meet the specific needs of the community. This section describes three alternative implementation approaches: (1) direct adoption of a specific evidence- based prevention program, (2) adaptation of an evidence-based interven- tion to community needs, and (3) community-driven implementation. Table 11-1 summarizes the advantages and disadvantages of each. These three approaches are not mutually exclusive or exhaustive of all potential approaches. Each requires an active partnership among community Âleaders, organizations and institutions, and researchers and must address issues of trust, power, priority, and action. The appropriate approach in a given com- munity will depend on its characteristics and priorities and the availability of an existing evidence-based program that matches its needs. Ideally, evaluation is a component of all three approaches to shed light on why a specific approach works in a particular community or how to generalize knowledge about successful implementation to other programs, communi- ties, or institutional settings. Adoption of an Existing Evidence-Based Program A communityâs adoption of a specific prevention program involves deliv- ering the program with high fidelity, increasing the likelihood that its impact will be similar to that found in the original studies. Typically, programs have met a specific standard of evidence, often articulated by federal, state, or other external funding sources (Halfors, Pankratz, and Hartman, 2007). Standardized curricula, teaching manuals, or taped media help deliver the program in a manner similar to that used by the original researchers. Gener- ally, there is limited adaptation of the program to the cultural or historical characteristics or the particular interests of the community. Sites typically need sufficient local capacity and resources and technical assistance from the program developers or other certified trainers to ensure fidelity, monitoring, supervision, and sustainability (Elliott and Mihalic, 2004). Both the Nurse-Family Partnership Program and Life Skills Train- ing, considered strong evidence-based programs backed by research findings
IMPLEMENTATION AND DISSEMINATION 299 TABLE 11-1â Comparison of Three Implementation Approaches Model Advantages Disadvantages Implementation High program fidelity Program may not fit community of an existing needs, strengths, or capacities evidence-based Relatively high likelihood of program achieving intended impact Real-world implementation may differ dramatically from the way Known resources and originally tested requirements for effective implementation Lack of ownership in the program Likely continued funding under Few evidence-based programs have federal and state supported the capacity to provide technical evidence-based prevention assistance and training An evidence-based program may not target outcomes relevant to community Adaptation of an Ownership and high support Key program components may be existing program from community and potentially modified, thereby reducing to meet high adoption outcomes community needs Program more relevant to ethnic, Essential program components not racial, or linguistic characteristics always evident of community Reasonably likely to achieve impact Community- Can develop high community Lengthy period to develop driven acceptance and ownership community awareness, common implementation vision, and program Potential for broader implementation across different Potential for ineffectiveness or organizations and institutions iatrogenic effects within the community Challenges in obtaining funding for Opportunity to empirically sustaining a unique program evaluate the outcomes of programs accepted by the community and use quality improvement methods to enhance outcomes over time
300 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS from multiple randomized trials in different types of communities, are being implemented in specific communities using this approach. There is some evidence that they are flexible enough to provide benefit across communities with diverse ethnic backgrounds (Botvin, Griffin, et al., 2001). However, it often takes decades of longitudinal follow-up for a pro- gram to be designated as evidence-based, and the original program may not address the current needs or priorities of communities. Research-based programs rarely can meet the triple challenges of maintaining an active research program, a successful marketing strategy, and a qualified technical assistance and training program. In addition, it may be difficult to repro- duce in the community the level of expertise of staff used to deliver the intervention in the original study. Finally, importing a program may result in a lack of ownership in the community, negatively affecting the ability to sustain the program over time. Given increasing evidence of the importance of community engagement and technical assistance, several models have been developed to help com- munities build the infrastructure needed to identify and implement specific evidence-based programs (see Box 11-1). For example, the Communities That Care (CTC) model leads a community through an assessment process to select specific evidence-based programs. The CTC model strongly discourages BOX 11-1 Models for Community Implementation of Evidence-Based Programs Communities That Care Communities That Care (CTC), a prevention system designed to reduce ado- lescent delinquency and substance use, was built as part of the Center for Sub- stance Abuse Prevention approach to effective implementation (see http://ncadi. samhsa.gov/features/ctc/resources.aspx). It provides a process for communities, through a community prevention board, to identify their prevention priorities and develop a profile of community risk and protective factors. The CTC logic model involves community-level training and technical assistance for three steps: (1) community adoption of a science-based prevention framework, (2) creation of a plan for changing outcomes through a menu of evidence-based programs that tar- get risk and protective factors identified by the community, and (3) implementation and evaluation of these programs using both process and outcome evaluations. Currently, there are 56 available programs that meet CTCâs required standard of evidence.
IMPLEMENTATION AND DISSEMINATION 301 BOX 11-1 Continued CTCâs theory of change hypothesizes that it takes two to five years to observe changes in prioritized risk factors and five or more years to observe effects on delin- quency or substance use. CTCâs data driven process is being evaluated in multiple steps. The first step, a five-year nonexperimental study with 40 incorporated towns, assessed the degree to which they reported using tested and effective programs. In the next phase, 24 of these communities who had not reported already using such programs agreed to be part of a large randomized community-level trial to test the CTC model (Hawkins, 2006). Early findings from these communities indicate that CTC has positive effects on targeted risk factors and delinquent behavior (Hawkins, Brown, et al., 2008) as well as alcohol use and binge drinking (Hawkins, Oesterle, et al., in press). Longer term follow-up is under way. PROmoting School-community-university Partnerships to Enhance Resilience Model The PROmoting School-community-university Partnerships to Enhance Resil- ience (PROSPER) model (Spoth, Greenberg, et al., 2004; Greenberg, Feinberg, et al., 2007) has devised a system aimed at broad implementation of evidence- based programs designed to support positive youth development and reduce early substance use delivered to rural areas with supports at the local, regional, and state levels. Underlying this system is the building of an infrastructure that supports local ownership and capacity building as well as leadership and institu- tional support (Spoth, Greenberg, et al., 2004). Three groups are involved in the PROSPER partnership model: (1) faculty from land grant universities and affiliated cooperative extension staff, (2) the elementary and secondary school systems, and (3) community agency providers of services for children and families, along with other community stakeholders. The partnership benefited from the existing training and technical assistance infrastructure provided by the Extension System and the U.S. Department of Edu- cationâs Safe and Drug-Free Schools (SDFS) Program. Because the prevention programs in PROSPER are delivered by local practitioners, it focuses on building strong support of the schoolâlocal community team, which chooses interventions and is responsible for their implementation. At the state level, researchers work with regional Extension Service prevention coordinators and coordinators from the SDFS Program. These regional coordinators then provide support to local teams of extension agents, elementary and secondary school faculties and staffs, and community interagency coalition members. The long-term goal is to provide infrastructure support as well as direct assistance to sustain effective, empirically based programs in communities. This implementation model has national implications, as the Extension Service has more than 9,600 local agents working in 3,150 counties across the United States. The Department of Education has multiple technical assistance centers that support efforts to adopt empirically supported programs that can reduce sub- stance abuse, violence, and other conduct problems in the schools. Furthermore, the SDFS Program currently has coordinators in many schools to facilitate the implementation of such research-based programs.
302 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS program adaptation, based on evidence that delivery of evidence-based pro- grams as designed is likely to lead to the most successful prevention efforts. Adaptation of an Existing Program to the Community Adaptation of programs focuses on concerns about community or cultural relevance. A community identifies an evidence-based program that matches its needs, values, and resources and modifies or adopts elements of the program to maximize community acceptance, implementation, and sustainability. Researchers often work in close collaboration with commu- nity leaders to find ways to integrate components of prevention programs in ways that are acceptable and meaningful to the community and to evaluate results. There is long-standing consensus that health promotion and preven- tion programs should be culturally sensitive, along with concerns about whether a given prevention intervention is generic enough to be efficacious and effective with diverse cultures (Resnicow, Baranowski, et al., 1999; Seto, 2001; Woods, Montgomery, and Herring, 2004; Weeks, Schensul, et al., 1995; Hutchinson and Cooney, 1998). Prevention programs must also be mindful of developmental processes, Â reinforcements of risk behavior, relevant contextual factors, and a populationâs unique risk profile (Brown, DiClemente, and Park, 1992). A few studies have shown that making adaptations to different cultural groups while maintaining core elements of programs implemented with fidelity can produce strong results across different cultural groups (Botvin, Schinke, et al., 1994; Botvin, Baker, et al., 1995; Botvin, Schinke, et al., 1995; Reid, Webster-Stratton, and Beauchaine, 2001). However, there is currently no consensus and limited scientific evidence on the key elements that determine the necessary balance between program adaptation and program fidelity. Bell, Bhana, and colleagues (2008) point out that, for an intervention to be culturally sensitive, it must have content that is welcoming to the target culture, contain issues of relevance to the culture, not be offensive, and be familiar to and endorsed by the culture. If a given intervention embodies generic principles of health behavior change, such as aspects that create social fabric, generate connectedness, help develop social skills, build self- esteem, facilitate some social monitoring, and help minimize trauma (Bell, Flay, and Paikoff, 2002), it can usually be adapted to have an appropriate level of cultural sensitivity (Bhana, Petersen, et al., 2004; Peterson, 2004; L Â aFromboise and Lewis, 2008; LaFromboise, 1995). For example, if going on a spirit quest builds self-esteem in American Indian culture, efforts to build self-esteem in American Indians might best be served by a spirit quest exercise instead of formation of a soccer team (Bell, 2005; see also
IMPLEMENTATION AND DISSEMINATION 303 BOX 11-2 A Program Adaptation for an American Indian Population An American Indian tribe in the Southwest worked in collaboration with aca- demic researchers to create the American Indian Life Skills (AILS) intervention for the purpose of reducing the factors associated with suicidal behavior Â(LaFromboise and Lewis, 2008). AILS was found to have a positive impact on American Indian high school studentsâ feelings of hopelessness, suicidal Â ideation, and ability to intervene in a peer suicidal crisis situation (LaFromboise and Â Howard-Pitney, 1993). When used as a comprehensive suicide prevention approach, the interven- tion demonstrated a substantial drop in suicidal gestures and attempts. Although suicide deaths neither declined significantly nor increased, the total number of self-destructive acts declined by 73 percent (May, Sena, et al., 2005). Extensive input was solicited from members of the tribe initiating AILS in order to fit its cultural norms. Key aspects of giving instruction, problem solv- ing, and helping others in that culture were examined. Focus groups members were selected by community leaders to give guidance on intervention content, implementation issues, and intervention refinement. It was believed that suicidal behavior could be attributed to direct modeling influences (e.g., peer or extended family memberâs suicidal behavior) in conjunction with environmental influences (e.g., geographic isolation) and individual characteristics (e.g., hopelessness, drug use) that mediate decisions related to self-destructive behavior. Life Skills Training was used throughout the intervention to complement t Âraditional ways of shaping behavior. Each skill-building activity was selected from research supporting best practices for social emotional regulation and cog- nitive skills development, including methods of group cognitive and behavioral treatment. Needed modifications were made to strategies identified. For example, in l Âessons on recognizing and overcoming depression, the Pleasant Events ÂSchedule (Lewinsohn, Munoz, et al., 1986) was modified to reflect American Indian adoles- cent socialization in the reservation context, renamed âDepression Busters,â and used as the basis for both an intervention activity and a homework assignment. Items such as âtalking on the telephoneâ or âplaying a musical instrumentâ were retained, while new items, such as âdoing heavy outdoor work (e.g., cutting or chopping wood, clearing land)â or âbeing at weddings and other ceremonies,â were added. In lessons addressing stress management, the eight ways of coping advanced by Folkman and Lazarus were shared in the focus groups to better determine cultural coping preferences and coping styles (Folkman, Lazarus, et al., 1986). The coping strategies most highly endorsed by participants in these groups were emphasized throughout the intervention. This âhybrid-likeâ approach (Castro, Barrera, and Martinez, 2004) encouraged the inclusion of traditional and contemporary tribal world views in the intervention without compromising its core psychological components. After several formative evaluations with diverse tribal groups, AILS has been refined to address the needs of both traditional and pan-tribal adolescents (LaFromboise, 1995; LaFromboise, Coleman, and Hernandez, 1991). It has been continued
304 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS BOX 11-2 Continued implemented by interventionists (including teachers) for work with urban and res- ervation youth during in-school, after-school, and community-based programs for American Indian youth. AILS is thought to be broad enough to address concerns across diverse American Indian tribal groups yet respectful of distinctive and heterogeneous cultural beliefs and practices. The program received support in 2007 from three suicide prevention projects, funded by the Substance Abuse and Mental Health Services Administration, to train American Indian interventionists on a wide-scale basis, to complete an early adolescent version of the intervention, and to create an implementation guide. Efforts to evaluate AILS in an urban Indian education program are currently under way. Box 11-2). Bernal, Bonilla, and Bellido (1995) provide a framework for developing culturally sensitive interventions that calls for consideration of language, persons, metaphors, content, concepts, goals, methods, and context. On the other hand, research has indicated that, although cultural or other adaptations made by practitioners that reduce dosage or eliminate critical core content can increase retention by up to 40 percent, they also reduce positive outcomes (Kumpfer, Alvarado, et al., 2002). For example, efforts to create and disseminate best-practice components of the Nurse- Family Partnership Program failed to produce the same results as the controlled trial replications (Alper, 2002; Olds, 2002). While research on dissemination of tested and effective prevention programs appears war- ranted, more research to identify the active ingredients of those programs is required before adaptation and dissemination of best practices distilled from these programs are warranted. In general, there has been a dearth of research on cultural, racial, and ethnic issues involved in interventions aimed at preventing mental, emo- tional, and behavioral (MEB) disorders (U.S. Public Health Service, 2001a) and even less research on the effectiveness of specific prevention strategies when implemented in a population other than that originally targeted by a trial. However, several models are being used to examine the extent to which program adaptation can be used to address the unique cultural needs of communities. Castro, Barrera, and Martinez (2004), for example, describe a hybrid approach to modifying the content and delivery of an existing prevention program. This area needs more research, as few empiri- cal Âstudies have examined alternative strategies. One method of enhancing cultural sensitivity and cultural relevance is
IMPLEMENTATION AND DISSEMINATION 305 to involve the community in every aspect of a prevention trial (LaFromboise and Lewis, 2008; LaFromboise, 1996; Madison, McKay, et al., 2000; McCormick, McKay, et al., 2000; Baptiste, Blachman, et al., 2007; Bell, Bhana, et al., 2007; McKay, Hibbert, et al., 2007; Pinto, McKay, et al., 2007). However, developing and maintaining community involvement throughout all stages of program implementation present considerable challenges, as discussed below. Community-Driven Implementation Community-driven implementation builds heavily on the decision mak- ing of community leaders, often in partnership with researchers, with a focus on improving the community relevance and sustainability of a program. Implementation is guided by a community-driven agenda and staged imple- mentation of a prevention program, in some cases including development, implementation, and testing of a locally developed intervention. Evidence- based programs or principles are often introduced by research partners relatively late in the process. Built on the community-based participatory research approach, an agenda for community action is developed through a cooperative process with community members and multiple community constituencies. The involvement of researchers in identifying priorities may be quite limited or very involved (Minkler, 2004), but it always focuses on community leadership and establishment of an organizational structure for building and sustaining one or more interventions (Baptiste, Blachman, et al., 2007). In many minority communities, there is a history of mistrust of outsiders, government agencies, or researchers in particular (Thomas and Quinn, 1991), which influences the degree to which Â researchers are involved in decision making (McKay, Hibbert, et al., 2007). The traditions of research, including reliance on planned research designs, multiple assessments, and legal consent documents, are often viewed negatively by communities. Thus, researchers may begin as outside advisers who listen to the goals and needs of communities, with a part- nership in the decision-making processes evolving over time. The wealth of practical experience and wisdom in community-based organizations may offer opportunities for communities to establish an empirical basis for interventions with strong community support through communityâ research partnerships. Collaborations with key community constituents can (1) enhance the relevance of research questions, (2) help develop research procedures that are acceptable to potential community participants from diverse cultures, (3) address challenges to conducting community-based research, (4) maximize the usefulness of research findings, and (5) fos- ter the development of community-based resources to sustain prevention funding beyond grant funding (Israel, Schulz, et al., 1998; Institute of
306 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Medicine, 1998; Schensul, 1999; Jensen, Hoagwood, and Trickett, 1999; Wandersman, 2003). Efforts to move from efficacy and effectiveness to full-scale implementation can and often do begin early by establishing such partnerships (Fixsen, Naoom, et al., 2005). A number of prevention specialists have called for the scientific study of communityâresearch partnerships (Chinman, Hannah, et al., 2005; Spoth and Greenberg, 2005; Trickett and Espino, 2004; Wandersman, 2003). The principles that guide such partnerships are clear and involve Â researchers developing win-win relationships with communities in their efforts to foster trust and mutual respect (see Madison, McKay, et al., 2000; Israel, Schulz, et al., 2003; ÂTrickett and Espino, 2004; Bell, Bhana, et al., 2007; McKay, Hibbert, et al., 2007; Pinto, McKay, et al., 2007). Researchers and com- munity collaborators should attempt to develop shared vision and mission, BOX 11-3 CHAMP: Collaborative HIV Adolescent Mental Health Program The Community Collaborative Board for the CHAMP project builds on the framework for an academicâcommunity collaborative approach to HIV/AIDS risk reduction with urban adolescents (McKay, Hibbert, et al., 2007). The mission was âif the community likes the program, the research staff will help the community find ways to continue the program on its ownâ (Madison, McKay, et al., 2000). The CHAMP Community Collaborative Board structure is characterized by moderate- to high-intensity collaboration (Hatch, Moss, et al., 1993). All of the CHAMP Family Programs use community representatives as liaisons between youth and families in need and prevention programs, as suggested by research (Koroloff, Elliott, et al., 1994; McKay and Paikoff, 2007). Community parent facili- tators, who had participated in the program themselves, are trained to reach out to their neighbors and invite them to learn more about the program. In addition to providing factual program information, they are also able to share personal, firsthand program experience. Community members also play a role in delivering the intervention, helping to address issues of cultural sensitivity and addressing research concerns of efficacy and effectiveness while preparing the community for dissemination. All of the intervention sessions are cofacilitated by a mental health intern/parent facilitator team. The team receives weekly joint training in program content; skills related to facilitation of child, parent, and multiple family groups; and issues related to prevention research and protection of human subjects, including confidentiality and mandated safety issues. Grant funding to enhance leadership development among community board members was secured to help pave the way for the community to take over the
IMPLEMENTATION AND DISSEMINATION 307 consensus on strategies, and synergy in execution and implementation (Senge, 1994). Resources should be openly discussed with community mem- bers, who should benefit from the resources as much as do the research- ers. Thus resources, both tangible (e.g., Âresearchers employing community members and partners providing facilities for programs) and intangible (e.g., partnersâ knowledge of participants and researchersâ knowledge of research methods), should be shared (Suarez-Balcazar, Davis, et al., 2003). Collaboration must also involve team training in which researchers learn community issues and community partners learn research issues. Early involvement of communities, power sharing, mutual respect, community benefit, and cultural sensitivity (Sullivan, Kone, et al., 2001) are needed to meet these challenges. Box 11-3 describes a program aimed at HIV/AIDS risk reduction that is built on such a collaborative model. intervention from the research team (Madison, McKay, and the CHAMP Col- laborative Board, 1998). Community support was hypothesized to facilitate wider dissemination of prevention messages and strategies (Galbraith, Ricardo, et al., 1996; Schenshul, 1999; Stevenson and White, 1994). The team believed that given the business skills necessary to run such programs, large community-based agen- cies might be more able than academic research teams to retain proven programs within their infrastructure, enhancing the likelihood that a specific program would be sustained over time (Galbraith, Ricardo, et al., 1996; Goark and McCall, 1996). Community leaders were also responsible for the day-to-day research operation (with consultation from university researchers) (McKay, Chasse, et al., 2004). CHAMP-Chicago and New York were also funded to study how to transfer an academic research project (based at the University of Illinois in Chicago and Mt. Sinai School of Medicine in New York), with both efficacy and effective- ness components. For example, in Chicago the program was transferred to a community-based organization (Habilitative Systems Inc., a social service agency in urban Chicago). Key elements of the framework for this 13-year transfer are (1) ensuring a good academic-agency fit, (2) early planning for sustainability, (3) building in continuous quality improvement, and (4) balancing program adaptation with fidelity (Baptiste, Blachman, et al., 2007). The experience implementing the CHAMP Program in Chicago and New York helped inform the 2001 CHAMP-South Africa research project. Based on its success, in 2007, the South African HIV prevention intervention obtained private foundation funding to serve 500 families, many of whom were in the control condi- tion during the study.
308 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS EXAMPLES OF EXPERIENCE WITH IMPLEMENTATION OF EXISTING PREVENTION PROGRAMS In all fields of health, including the prevention of MEB disorders, there is a time lag between documentation that an intervention, program, policy, or practice improves health in a defined community and successful adop- tion of that program in society (Walker, Grimshaw, et al., 2003; Walker, Seeley, et al., 2008). Levels of implementation of preventive interventions are rarely measured; however, more information is available on implemen- tation of substance abuse prevention in schools than on other prevention interventions. In a national study of middle schools, Ringwalt, Ennett, and colleagues (2002) found that, while 81.8 percent of public and private schools offered a substance abuse prevention curriculum, only 26.8 percent were using 1 of 10 tested and effective curricula. Furthermore, even when schools and communities use tested and effective programs, they often fail to implement them with fidelity to the standards delineated by program designers (Ennett, Ringwalt, et al., 2003; Mitchell, Florin, and Stevenson, 2002; Wandersman and Florin, 2003). The National Study of Delinquency Prevention in Schools Â(Gottfredson and Gottfredson, 2002) found that only half of drug prevention curri- cula and one-fourth of mentoring programs met âdosageâ requirements (amount of studentsâ exposure to the subject). The rest delivered fewer and less frequent sessions than were specified by program developers. Moreover, only half of the programs were taught in accordance with the recommended methods of instruction. Hallfors and Godette (2002) report that only 19 percent of all surveyed school districts faithfully implemented tested and effective prevention cur- ricula. Similarly, Ennett, Ringwalt, and colleagues (2003) found that only 14 percent of middle school teachers of drug prevention curricula exposed their students to adequate content and means of delivery. Yet adherence to core program components is important to ensure outcomes. There is evi- dence that some tested and effective prevention programs work only when implemented with a high degree of fidelity (Abbott, OâDonnell, et al., 1998; Botvin, Mihalic, and Grotpeter, 1998; Henggeler, Brondino, et al., 1997; Kam, Greenberg, and Wells, 2003; Olweus, Limber, and Mihalic, 1999). M Â ethods for widespread dissemination of tested and effective prevention policies and programs with high levels of fidelity are needed Â (Farrington and Welsh, 2006; Spoth and Greenberg, 2005; Wandersman and Florin, 2003). Often, federal or state guidelines dictate reimbursement only for using approved or evidence-based programs. For example, the tobacco settlement money from several states allows for funding of specifically named preven- tion programs. Colorado amended its revised statutes to provide continued
IMPLEMENTATION AND DISSEMINATION 309 funding for the Colorado Nurse-Family Partnership Program. Recent con- gressional action related to child abuse programs sets aside funds for com- petitive state grants to implement evidence-based home visitation models. Numerous guidelines for best practices in program implementation have been published at the federal and state levels, although somewhat different, and sometimes confusing, criteria have been used for the selection of recom- mended programs across reviews and federal agencies (see Chapter 12). The Blueprints for Violence Prevention project at the University of Colorado, originally stimulated with federal funds from the U.S. Department of Justice, was institutionalized as an effort to facilitate community-level adoption of specific violence prevention interventions. Similarly, the Safe Schools Healthy Students Initiative, a joint program of three federal agencies, encourages the use of evidence-based programs. However, empirical evidence of the imple- mentation experiences or results of this program are lacking. One of the most promising approaches to prevention involves early education of children to prepare them for the challenges of reading and learning, as well as to help them develop the social skills necessary to develop positive relations with their peers as well as adults. Yet far fewer children in the United States ages 3-5 receive preschool or center-based early education than is the case in many other developed countries (Yosikawah, Schindler, and Caronongen, 2007). Children from poor families and chil- dren with less educated mothers are dramatically less likely to receive these services compared with those with higher incomes and mothers with more education (U.S. Department of Education, 2007). Although there are exceptions, overall implementation of effective interventions has been modest at best. Scientific evidence of health benefits and standard methods of dissemination alone appear to work only under limited circumstances. Successful examples of implementation that do exist across the life span and across different settings are reviewed below. Implementation of Prevention Programs in Early Childhood Chapter 6 outlines a number of early childhood programs that have demonstrated positive outcomes for children and their families. The history of one promising program implemented in multiple states provides impor- tant lessons on the significance of implementation fidelity and program design. Healthy Families America (HFA) is a program aiming to prevent child abuse and neglect that was modeled after the Hawaii Healthy Start program and implemented state-wide in several states. The core of this pro- gram involves identifying parents at high risk of abuse and neglect of their children. Parents are identified through broad-based screening and then offered voluntary home visiting services delivered by paraprofessionals for a period of three to five years (Duggan, McFarlane, et al., 2004). A study
310 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS conducted in Hawaii yielded disappointing results, with as many negative as positive impacts on key family process outcomes (Duggan, McFarlane, et al., 2004). The evaluators offered several potential explanations for these results: The program may have been poorly implemented, as half (51 percent) of the parents dropped out within the first year and participating families received fewer home visits than intended; the paraprofessional staff may not have had sufficient skills to identify high-risk situations and engage parents in the process of reducing risks associated with abusive parenting; and a shift away from an emphasis on recognizing and addressing risks for abusive parenting toward an early intervention philosophy of parent- driven goal setting, which was caused by funding requirements, may have compromised its effectiveness. A recent evaluation of an augmented HFA program, with a sharper focus on using cognitive appraisal theory to reduce risks for abuse and neglect, as well as better implementation practices, yielded considerably more favorable results compared with both the unenhanced HFA pro- gram and a control group that did not receive any home visiting services (Bugental, Ellerson, et al., 2002). These positive findings were particularly evident for medically vulnerable infants, such as those born prematurely or those with low Apgar scores (assessing physical condition after delivery) at birth. Although the study was small and thus in need of replication, this finding illustrates the current effort among the nationâs largest home visiting models to use evaluation findings to promote program improvement. The lessons learned from this study (i.e., the importance of engaging families, providing high-quality training and ongoing supervision of staff, and ensur- ing consistent and well-implemented service delivery) illustrate the value of program accountability as a strategy for continuous enhancement rather than as a vehicle for terminating potentially effective services that produce initially disappointing results. Similarly, lessons learned from the Healthy Families New York evaluation (see Chapter 6) reinforce the likely value of targeting first-time mothers with limited resources. The implementation of quality early childhood programs at scale con- tinues to be a vexing problem in the field. Few evaluations have linked implementation quality to the magnitude of impacts on childrenâs social- emotional or mental health outcomes. Research from Early Head Startâs 17-site evaluation has shown that sites with earlier and more complete implementation, as measured by the Early Head Start Program Performance Standards, had stronger positive impacts on child social-emotional and cognitive outcomes than those with later or less complete implementation (Love, Kisker, et al., 2002). Other research on quality in at-scale programs has found that Head Start classrooms in general are in the âgoodâ (though not âexcellentâ) range of quality on the Early Childhood Environment Rat- ings Scale (Administration for Children and Families, 2006). Research on
IMPLEMENTATION AND DISSEMINATION 311 the impacts of thresholds of quality in this and other scales measuring early childhood care and education environments is currently lacking. Nursery and preschool programs represent a clear opportunity to intervene early in the lives of young children. Implementation of Prevention Programs in Schools The implementation of universal school-based interventions faces con- siderable challenges, including access to and approval of schools that are often overburdened with other academic and policy-related priorities. Mul- tiple levels of approval from superintendents, principals, teachers, and school boards and community partners may also be required. In addition, the relatively low dosage that is common to most universal strategies may not offer sufficient exposure to impact children already at very high risk for a specific disorder (e.g., depression or anxiety). Universal interventions are likely to be most effective when they have sufficient duration and intensity, target the development of protective factors and resilience likely to impact risk for multiple disorders, or target problems common to a large segment of the population (e.g., bullying, early alcohol use). Nevertheless, there are several opportunities for successful implemen- tation in schools, particularly since there has been a substantial change in legislation focused on evidence-based prevention at both the federal and state levels. Two dramatic examples are the Safe and Drug-Free Schools Act of 1999, which states âprinciples of effectiveness,â and the No Child Left Behind Act of 2001 (NCLB), which calls for school districts to implement evidence-based programming (Hallfors and Godette, 2002). Other oppor- tunities have been created by several state legislatures that have mandated the use of character education in schools. Prevention programs relevant to schools focus on the school and its structure (organizational features, school rules), classroom behavior man- agement, and curricula that teach students new skills, recently termed social-emotional learning (see Chapters 6 and 7). Some form of curricula- oriented prevention programs for substance abuse is in a large number of the nationâs middle schools. Schools have also become a venue for both targeted and indicated interventions. As an example, Title I Part A provides additional funding to schools that have poverty levels above 40 percent, and such programs can be delivered either universally or in a targeted fashion to those at higher risk. Some school-based indicated interventions focus on children who are showing early indications of a potential disorder (e.g., high rates of aggres- sive behavior, anxiety, depression, or other forms of maladjustment). These interventions are also usually provided in special groups, and most support the development of social and emotional learning skills. For some preven-
312 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS tion programs, school is only one of several components that also may include family or community-based interventions. However, given the extra costs associated with multicomponent interventions, they are more often used with either selective or indicated models (e.g., the Incredible Years; Webster-ÂStratton, 1998), or with universal models that occur in high-risk neighborhood schools (e.g., Seattle Social Development project; Hawkins, Kosterman, et al., 2005). Implementation of Prevention in Child Welfare and Juvenile Justice Settings In one study of scale-up of an intervention to help foster parents improve their childrenâs emotional and behavioral functioning (Multi- dimensional Treatment Foster Care), a âcascading training modelâ was used and tested for effectiveness. After an initial phase representing an efficacy trial, a second phase of treatment for foster parents was deliv- ered by Â paraprofessionals, who were intensively overseen by an onsite supervisor and an experienced clinical consultant. During a third phase, the Â paraprofessionals trained a second cohort of staff. In this phase, the clinical consultants oversaw the supervision but did not have any direct contact with the new staff. The evaluation found that both phases resulted in decreases in childrenâs problem behaviors, with no significant difference between the two phases. This suggests that intensive training and super- vision can enable âthird-generationâ staff to scale up and implement an intervention with fidelity (Price, Chamberlain, et al., 2008). Implementation of Prevention in Primary Care Settings Few preventive interventions have been tested in primary care set- tings, although collaborative treatment models involving primary care and behavioral health staff have begun to emerge (Forrest, Glade, et al., 1999; G Â uevara, Rothbard, et al., 2007), and physicians should routinely screen for behavioral and developmental concerns (see Chapter 8). Pediatric primary care settings are seeing significant numbers of patients with mental health problems (Horowitz, Leaf, et al., 1992; Briggs-Gown, Horwitz, et al., 2000; Kelleher, McInerney, et al., 2000), with some estimates that the number of office visits for mental health problems has increased by 2.5 (Kelleher, McInerney, et al., 2000; Zito, Safer, et al., 1999). One promising primary care intervention involves a strategy to encourage teens to use a primary care Internet-based intervention to prevent depression (Van ÂVoorhees, Ellis, et al., 2007). As outlined in Chapter 8, primary care settings represent a significant opportunity for development of new approaches to identify and respond to parentsâ concerns about their childrenâs behavioral and
IMPLEMENTATION AND DISSEMINATION 313 emotional health. Such approaches would facilitate coordination of all the childrenâs health care needs and reinforce the integral nature of physical and mental health care needs. IMPLEMENTATION CHALLENGES Fixsen, Naoom, and colleagues (2005) note that â[successful] imple- mentation is synonymous with coordinated change at system, organization, program, and practice levels.â This coordination is not easy to achieve, and indeed these authors note that poor implementation of a beneficial program can come from unsupportive policies or regulations or a lack of funding opportunities at the federal, state, or local level; a lack of organiza- tional commitment, capacity, or leadership; poorly chosen or high turnover among intervention agents or practitioners; or a lack of involvement in or ownership in the program by the community. Until recently there has been little support from the federal or state governments for prevention activities and even less for building an infrastructure that facilitates these efforts. It is also often difficult to sustain attention on a specific problem, as evidenced by the current lowering of priorities for HIV prevention and youth tobacco prevention (U.S. Government Accountability Office, 2007; Institute of Medicine, 2006b). Coordinating all these issues would be difficult enough if policy Âmakers, organizations, practitioners, community leaders, and consumers all spoke the same language and shared a common vision. However, these groups often have vastly different world views and priorities and are often reluc- tant to learn about each othersâ perspectives. The previous sections have described some key challenges to the implementation of preventive interventions, including the need to bal- ance cultural adaptation and fidelity, the difficulty of forming essential c Â ommunity-research partnerships, and the time lag between documentation that an implementation is effective and its successful adoption. This sec- tion reviews additional implementation challenges: funding; service system priorities; training, monitoring, and capacity building; data systems; low participation and retention rates; and organizational context. Funding Obtaining adequate funding is a challenge for all types of implemen- tation. Program cost, including the cost of labor, materials, and technical assistance, is often just as or even more important to communities and policy makers than effectiveness. Yet not only are evaluations of the benefits and costs of prevention programs relatively uncommon (see Chapter 9), but also collection of cost data is as well. In one analysis of prevention pro-
314 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS grams targeting children younger than age 5, cost data were available for only a small percentage of the studies (Brown, Berndt, et al., 2000). There have been limited commitment and funding for prevention train- ing and materials at government and community levels (Hallfors, Pankratz, and Hartman, 2007), with few targeted funding sources and available sources providing limited amounts. For example, SDFS, which represents 70 percent of school funding for drug prevention programs, provides an average of only $6.30 per child Â(Hallfors, Pankratz, and Hartman, 2007), well below the copying costs let alone the costs of training and sustaining an effective prevention program. FurtherÂmore, two-thirds of the states do not provide any additional funding beyond that provided by SDFS. A report on Nevadaâs implementation of school-based substance abuse and violence prevention programs concluded that funding was inadequate in most school districts to implement the type of prevention program needed (Nevada State Department of Education, 1998). Half of the states have two or fewer full-time staff available to support schoolsâ selection and implementation of drug prevention programs statewide Â(Hallfors, Pankratz, and Hartman, 2007). The Safe Schools/Healthy Students Grant Program, funded jointly by three federal agencies, aims to help local communities develop integrated programming that involves prevention, treatment, and school reform efforts in K-12 based on evidence-based interventions. However, although over 150 communities have been funded through this program, there is as yet little published research on this model. Service System Priorities Prevention is often tangential or only weakly related to the mission of the institutions and communities in which its programs could be housed, leading to limited infrastructure to support and sustain prevention pro- grams in their natural settings (Greenberg, 2004; Spoth, Greenberg, et al., 2004). For example, integrating mental health and drug prevention into the existing primary health care system would require more accessible reimbursement mechanisms. The mental health system, which is primarily focused on treatment of disorder, would need to be reoriented. Similarly, the primary mission of schools is to educate students, with an emphasis on core subjects, like science, math, history, and reading; it is unlikely that they will support a program that does not directly relate to this primary mission (Kellam, 2000). American schools face many com- peting demands, and education leaders must make difficult choices about priorities (Adelman and Taylor, 2000; Berends, Bodilly, and Kirby, 2002; Hall and Hord, 2001). Currently most education leaders focus on the stu- dent academic performance requirements of the NCLB. While a potential
IMPLEMENTATION AND DISSEMINATION 315 benefit of the act is the promotion of academic excellence and equity, it has also led to high-stakes testing, a narrowing of the curriculum, and a loss of the âwhole childâ in education. Critics caution that, without attending to studentsâ social and emotional needs, many of these actions may be Â ineffective at best and harmful at worst, especially for economi- cally disadvantaged groups (Meier and Woods, 2004). A consequence of NCLB has been a marginalization of most prevention efforts, as they have not been well linked to educational outcomes (especially achievement test scores). Few prevention programs take the time or effort to integrate into the schoolsâ central mission (Smith, Swisher, et al., 2004). In fact, few even assess academic performance as an outcome (Hoagwood, Olin, et al., 2007; Durlak, Weissberg, et al., 2007), although there is some indication that programs focused on social and emotional learning can increase academic achievement (Durlak, Weissberg, et al., 2007). As a result of federal and state legislation, U.S. schools are rapidly reorganizÂing and striving to develop broader and more comprehensive m Â odels of reform that use clear goals, standards, and benchmarks for out- comes (Education Commission of the States, 2001; Togerni and ÂAnderson, 2003). As empirically validated programs have accumulated and been increasingly adopted (Ringwalt, Ennett, et al., 2002), schools are searching for integrated models with a clear scope and sequence from prekindergar- ten through grade 12 (Collaborative for Academic, Social, and Emotional Learning, 2003; Elias, Zins, et al., 1997). Current evidence-based programs span relatively small parts of this age span. As the process of school reform grows, researchers and practitioners will need to work together to develop pre-K-12 guidelines and consider how all the elements of evidence-based programs and policies fit together in the context of an overall schoolwide or school district effort, how they increase studentsâ school success, and how to ensure that coordinated, multiyear programs will be implemented effectively (Adelman and Taylor, 2000; Osher, Dwyer, and Jackson, 2002). Integration of program mod- els across developmental periods, long-term curricular planning involv- ing both Â researchers and practitioners, adequate local infrastructure to support prevention activities, teacher training and technical assistance, and appropriate evaluation of process and outcome must be part of this process (Greenberg, Weissberg, et al., 2003). Legislation in both Illinois and New York now requires that schools develop plans for social and emotional learning, models for learning standards, and benchmarks across all grade levels. Schoolâcommunity systems with programming integrated between uni- versal, classroom programs and either selective prevention (e.g., counselor- led programs with students who are experiencing divorce, bereavement,
316 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS or other trauma), indicated prevention (for children identified as having aggression, peer problems, prodromal signs of depression, etc.), or treat- ment are rare. The fragmented nature of the models created by curricu- lum developers and researchers as well as the often fragmented planning between schools, government agencies, and the private sector of human services contribute to this problem. Social settings that have not been often targeted to house preventive interventions include important microsystems and exosystems (systems that affect the child but do not directly include the child, such as a parentâs workplace) of human development (Bronfenbrenner and Morris, 1998). Workforce development organizations, for example, have the potential to influence parental employment, which in turn is related to early child- hood development. Community-based organizations (aside from child care and preschool programs) have also been underutilized. They may be set- tings that families trust (e.g., organizations serving immigrant populations) and that therefore may be productive settings for child- or family-focused programs. The primary health care system has also generally been over- looked as a setting for preventive interventions. Finally, the bulk of child careâbased interventions have occurred in center-based settings, with family day care settings rarely targeted for quality improvement or implementation of specific preventive curricula. Other service systems also have potentially much to gain from preven- tion. Communities have to invest significant resources to handle delinquent youth through the juvenile justice system, to counter ineffective or unsafe parenting through the foster care system, and to counter difficulties in learning or behavior through special education. The missions of all these programs are clear: to provide services to those who are in serious need. However, these systems usually do not embrace the mission of preventing these problems from arising in the first place. Unless prevention can find ways to integrate its work into the central missions of these and other com- munity institutions (Kellam, 2000), the prevention focus will continue to be lacking. Two reports of the Institute of Medicine call for an increase in b Â idirectional communication between researchers and organizations and social service settings in which prevention can be housed (Institute of Medi- cine, 1994, 1998). It seems sensible that research efforts should be directed toward understanding and facilitating such communications, although little research has been conducted in this area. One study, surveying both researchers and practitioners who attended the same bereavement confer- ence on modes of communication, found relatively modest overlap (Bridg- ing Work Group, 2005).
IMPLEMENTATION AND DISSEMINATION 317 Training, Monitoring, and Capacity Building A review by Mihalic and Irwin (2003) concluded that a consistently important factor in the success or failure of implementation of evidence- based interventions is the quality of ongoing technical assistance. With appropriate training and monitoring, programs can be disseminated with fidelity (Fagan and Mihalic, 2003; Spoth, Guyll, et al., 2007; Spoth, Redmond, et al., 2007). However, systems for delivering proactive technical assistance are limited and generally not up to the task required for large- scale dissemination (Mitchell, Florin, and Stevenson, 2002). For example, the Blueprints project attempted to implement a set of 10 empirically sup- ported programs, but only 4 of these programs had sufficient organizational capacity to implement the intervention in 10 different communities per year (Elliott and Mihalic, 2004). The Blueprints project has a rigorous system for identifying violence prevention programs with a very high level of evi- dence, facilitates tests of replication of promising programs, disseminates knowledge of these programs to communities, and provides technical sup- port for community implementation, with the direct involvement of the programâs developers. One of its key findings was that both the program developers and the implementation sites often required substantial multi- year technical assistance. For schools, the major training difficulties were lack of time to work with the model and the need for continual training due to staff turnover (Elliott and Mihalic, 2004). With intensive effort, however, sites were able to implement chosen programs with high fidelity, often approaching or exceeding the level of fidelity achieved by the program developers in the original study. The Blueprints project examined the degree to which chosen programs were adapted locally, even though most of the intervention trainers discour- aged adaptation. Very little local adaptation in these particular programs was needed to achieve acceptance, participation, and quality implemen- tation, even in diverse communities, and fidelity and sustainability were emphasized. However, by design, Blueprints does not evaluate impact on behavioral outcomes in the implementation communities, relying instead on each programâs earlier empirical success. It is possible that the programs had different rates of success in the implementation sites compared with those in the original studies. To address capacity-building and training needs, the Nurse-Family Partnership has established a nonprofit National Service Office to develop community capacity to implement a home visitation program with high fidelity and provide training and technical assistance (Olds, Hill, et al., 2003). In addition to training and technical assistance provided by program
318 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS developers, other organizations, including federally funded technical assis- tance centers (see Box 12-2) and state-level organizations Â (Pennsylvania Commission on Crime and Delinquency; Neal, Altman and Burritt, 2003; New York State Office of Mental Health) are providing assistance in implementing prevention programs, particularly substance abuse and vio- lence prevention programs. In addition, programs are being encouraged to provide manuals and other materials to assist in the implementation of programs. Data Systems Data systems that integrate family, school, and developmental informa- tion could be a useful tool for targeting and monitoring prevention pro- grams. NCLB legislation has provided an as yet unattained opportunity to use academic information to inform prevention needs. Integration of service and data systems for early childhood health, learning, and mental health, for example, could reduce duplication of services, track families across systems, identify children and families who are particularly vulnerable, link family need levels to services, and assess delivery and outcomes for diverse families in particular communities (Knitzer and Lefkowitz, 2006; Schorr and Marchand, 2007). Such integration has recently become a policy focus in the early child- hood field. For example, the Maternal and Child Health Bureauâs Early Childhood Comprehensive Systems initiative funds states (currently 47) to help coordinate services related to early health care, education, mental health, and family support (Johnson and Theberge, 2007). Studies in other state policy areas (e.g., state-level expenditures on state prekindergarten related to childrenâs cognitive and social-emotional outcomes in the Early Childhood Longitudinal Study-Kindergarten Cohort; Magnuson, Ruhm, and Waldfogel, 2007) could be applied to future studies on state variation in early childhood policies, such as systems coordination. Controlling for other unobserved state policy characteristics, such as concurrent policy change in other areas, is a challenge. In general, evaluation and quality improvement approaches require adequate data collection, storage, and analysis. Low Participation and Retention Rates An intervention may be poorly implemented due to a communityâs being overly optimistic about its capacity to provide the intervention to â See http://prevention.psu.edu/project/delinquencyandviolenceprevention.html. â See http://www.omh.state.ny.us/omhweb/sv/schlviol.htm.
IMPLEMENTATION AND DISSEMINATION 319 sufficient numbers of the target population (Chinman, Hannah, et al., 2005; Miller and Shinn, 2005). Brown and Liao (1999) note that âeven well- designed, efficacious interventions may fail when they are not delivered or implemented at full strength.â These authors opine that an intervention may not succeed if a high level of participation cannot be sustained throughout the intervention period. They note, as well, that an interventionâs benefit vanishes if there is low participation or if the intervention is not delivered to those who are likely to benefit from it. Individual choice to participate is not a major factor in the adoption of universal classroom or school pre- vention programs by a school because they generally involve all students. The same is not true of programs that require individuals to choose to par- ticipate, including most selective or indicated prevention programs. Often, when one of these prevention programs administered individually or by a small group is available in a community, proportionally few families opt to participate in it (Flay, Biglan, et al., 2005). When there is low individual-level participation, the overall benefit of the program in the community will typically also be low (population-level benefit is the proportion who participate times the effect size for the par- ticipants) (Braver and Smith, 1996; Brown and Liao, 1999). The overall effectiveness of family-based prevention programs (Reid, Webster-Stratton, and Hammond, 2003; Spoth and Redmond, 2000; Epstein, 1991; Eccles and Harold, 1996; Sheldon, 2003; Ialongo, Werthamer, et al., 1999) is particularly affected by low participation levels because families often make individual choices to participate, both initially and over time. For example, a significant number of families enrolled in home visita- tion programs drop out over the course of two to three years. Ialongo, Werthamer, and colleagues (1999) had reasonably high population-based participation rates (35 percent of all first grade parents attended six of the seven sessions, 13 percent attended none of the sessions), with fam- ilies on average receiving half the intervention, and positive long-term results. ÂFamily-based programs focusing on stressful events, such as divorce ( Â Wolchik, Sandler, et al., 2002; Forgatch and DeGarmo, 1999) or bereave- ment (Sandler, Ayers, et al., 2003), as well as community-wide parent train- ing (Brody, Murray, 2006b), also require extensive community engagement and recruitment to get acceptable participation levels. A critical concern for family-based programs is to increase participation, particularly for the subset at high risk that could most benefit (Brown and Liao, 1999). In addi- tion, attention is needed for strategies for increasing participation rates in real-world contexts that may not be able to offer the same incentives pos- sible in the research context. Lack of cultural relevance may contribute to low participation. How- ever, universal family prevention programs designed specifically for specific minority groups (Brody, Murry, 2006b; Prado, Pantin, et al., 2007) have
320 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS also experienced significant recruitment challenges. The use of focus groups and community partnerships to adapt the intervention and select settings for data collection and intervention, as well as inclusion of program and evaluation staff with similar cultural backgrounds, have allowed these stud- ies to maintain excellent participation rates over time (Murry and Brody, 2004). Other factors may also contribute to low participation rates, includ- ing the stigma associated with a program aimed at mental health, substance use, or problem behaviors; competing family demands, including multiple jobs or shift work; and community distrust of researchers. Organizational Context Successful implementation, including the ability to sustain a program, requires investments in people, relationships, and time, as well as coordi- nation around such critical issues as staffing and funding (Neumann and Sogolow, 2000). Organizations with little flexibility, fewer connections with professionals, and limited history with innovation may be the most in need of change but least capable of achieving it (Rogers, 1995; Kondrat, Greene, and Winbush, 2002; Hoagwood, Burns, et al., 2001; Schoenwald and ÂHoagwood, 2001). Many agencies that lack academic or other affili- ations often need technical assistance, staff support, and implementation resources to implement preventive interventions (Spoth, Kavanagh, and Dishion, 2002). Community-based agencies often grapple with high staff turnover and lack of adequate space, facilities, and equipment, all of which can interfere with sustainability (Kellam and Langevin, 2003; Swisher, 2000). Technical tasks, such as maintenance of data systems to assess risk factors, fidelity, outcomes, and satisfaction, are especially challenging when implementation is guided by community-based organizations or partner- ships (Dzewaltowski, Estabrooks, et al., 2004). Empirical work demonstrates that the types of changes required to implement and sustain preventive interventions are difficult to achieve. Robertson, Roberts, and Porras (1993) noted in a meta-analysis that orga- nizations that aimed at changing factors, such as organizational climate and culture, along with technological and strategic factors, were more successful than organizations that targeted only one of these areas. Valenteâs (1996) social network threshold model for innovation, which pairs local champi- ons of a programâwho can provide information outside the organization about a programâwith those in an organization who can change or shape its agenda is relevant to implementation of prevention programs. Both organizational climate and culture reflect the norms, expectations, and values of the organization, and they have strong influences on inno- vation and the adoption of new mental health service programs (Glisson and James, 2002). Glisson and colleagues use a change agent to facilitate
IMPLEMENTATION AND DISSEMINATION 321 stakeholder support in a program, external relationships with the organi- zation, and compatibility between the culture and the innovative program (Glisson, 2002; Glisson, Dukes, and Green, 2006). Because such organiza- tional change approaches have been successful in delivering mental health services, it may be a useful model to promote the adoption of prevention services as well. STRATEGIES THAT COMPLEMENT THE IMPLEMENTATION OF EVIDENCE-BASED INTERVENTIONS Despite the potential for preventing MEB disorders through the imple- mentation of evidence-based programs, there are limitations to relying exclusively on this approach. First, as noted above, evidence remains lim- ited about the ability of existing evidence-based programs to be effective for populations other than those that participated in the original evaluations of these programs. Moreover, implementation may be hampered if the original evidence was limited to efficacy trials rather than scientific evaluation based on characteristics as close to real-world conditions as possible. In addition, the widespread adoption, implementation, and maintenance of evidence- based programs will require a significant public investment. This section delineates additional strategies that can complement the implementation of evidence-based programs. Public Education Nationwide efforts to reduce cigarette smoking (Biglan and Taylor, 2000; Institute of Medicine, 2007b), one of the most successful public health efforts of the 20th century, illustrate the potential of public educa- tion strategies. The prevalence of smoking among adults has decreased by 58.2 percent since 1964, and smoking initiation by adolescent and young adults has also decreased (Institute of Medicine, 2007b). Although mul- tiple factors contributed to this remarkable decline, one contributor is the information that has been communicated to the public about the harm of tobacco use. At least two levels of government have adopted information and educa- tion strategies. At the local level, many school boards have required smok- ing prevention as a component of health education programs. At the federal level, Congress has mandated the publication of regular surgeon generalâs reports on smoking and health and, since 1966, has required warning labels on cigarette packs. The first surgeon generalâs report on smoking (U.S. Department of Health, Education, and Welfare, 1964) had an immediate and profound impact: In the first three months after its issuance, cigarette consumption fell by 15 percent. During a brief period, from 1967 to 1970,
322 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS the public was exposed to mandated antiÂsmoking commercials on televi- sion, which produced the first four-year decline in per capita smoking in the history of cigarette smoking. Subsequently, the Public Health Cigarette Smoking Act of 1969 banned cigarette ads on TV and radio effective in 1971 (Warner, 2006). Experience with antismoking social marketing campaigns in both Cali- fornia and Massachusetts has demonstrated the ability of professionally designed, well-funded, and sustained counteradvertising to decrease smok- ing. More recently, the highly acclaimed âtruth campaignâ antismoking ads produced by the American Legacy Foundation have been demonstrated to decrease smoking among youth (Farrelly, Davies, et al., 2006). We conclude that properly produced, financed, and distributed media campaigns can discourage youth smoking and reduce smoking among adults. Information and education clearly jump-started the antismoking cam- paign. All told, it seems highly likely that the combination of information and education interventions in the first decade of the antismoking campaign played a critical role in reducing youth smoking. MEB disorders and related problems are more diverse and complex than the single behavior of smoking. Nonetheless, there is an important potential role for public communication in terms of reducing the stigma associated with MEB disorders, conveying messages about the support structure needed to facilitate healthy development, and generating public support for relevant policies and principles as well as positive attitudes about the potential of the nationâs young people. Dissemination and Adoption of Common Principles As documented in this report, there are interconnections among MEB disorders, and the factors that contribute to them are interconnected. Simi- larly, there are common principles across a range of prevention approaches. Widespread communication of these principles to parents, community deci- sion makers, and policy makers could influence individual and collec- tive decisions supportive of common prevention practices. As discussed in Chapter 7, a number of common aspects of effective preventive inter- ventions as well as general lifestyle factors promote physical and mental health. Dissemination and adoption of these principles can contribute to the healthy development of the nationâs young people and the prevention of MEB disorders. Specifically: â¢ Effective preventive interventions reduce young peopleâs exposure to biologically and psychologically toxic events, such as harsh dis- cipline, abuse, and neglect. â¢ A common feature of most validated prevention programs is an
IMPLEMENTATION AND DISSEMINATION 323 emphasis on supportive environments or ânurturanceâ and positive reinforcement for prosocial behavior. â¢ Acceptance and encouragement in family, school, and community environments are more effective and desirable than confrontation or coercion. â¢ Such techniques as praise notes, peer-to-peer tutoring, and caregiver training can help facilitate the creation of nurturing environments. â¢ Adequate sleep, diet, and exercise, and television viewing limits can contribute to positive health outcomes. Principles such as these can be adopted in home, school, and commu- nity environments and need not be attached to specific prevention programs. Communication of the importance of these principles can reinforce desir- able behavior, minimize aversiveness, contribute to healthy development, and help promote a societal norm supportive of positive development. As with any interventions, empirical evaluation is also needed on how this information can be communicated to parents, teachers, caregivers, policy makers, and prevention practitioners and whether the communication of such information can have the same benefits that communications about smoking have had in reducing tobacco use. Public Policy Public policy changes made a significant contribution to the success of the tobacco control movement (Institute of Medicine, 2007b). Based on growing evidence of the harm of environmental tobacco smoke, tobacco control advocates have been able to push for local and state laws restrict- ing smoking. Smoke-free laws dramatically reduce workplace exposure to noxious chemicals. Such laws also lead to reductions in smoking among w Â orkers in the affected establishments. Thus, workers in smoke-free work- places are 3.8 percent more likely to quit smoking than are workers in workplaces that are not smoke free. Continuing smokers working in smoke- free environments reduce their daily cigarette consumption by an average of 3.1 cigarettes (Fichtenberg and Glantz, 2002). Public policy has also been a significant contributor to reductions in alcohol use and abuse. Taxation of beer, which increases its price, has been shown to reduce alcohol consumption among young people, especially those who are heavy drinkers (Biglan, Brennan, et al., 2004). Increasing the drinking age from 18 to 21 also has a well-documented impact on alcohol- related auto crash fatalities (Wagenaar and Toomey, 2002). The National Highway ÂTraffic Safety Administration estimated that increasing the drink- ing age from 18 to 21 saved 17,359 lives between 1975 and 1997. Pentz, Jasuja, and colleagues (2006) similarly argues that policies related to drug
324 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS use, including national media attention, have contributed to the ebb and flow of drug use among adolescents. States and the federal government have also enacted numerous laws to protect children from injury (e.g., car seat and bike helmet laws) and disease (e.g., requiring immunizations). Infant seat restraints in cars are now used by the majority of young children nationwide, and bike helmets are now common for children when just years ago they were only seen in bicycle races. Immunization protocols are well established, and proof of relevant immunizations is often required as a condition of school enrollment. Public policy could play a more significant role in mental health promo- tion and the prevention of MEB disorders. Given the relationship between poverty and MEB disorders (Conger, Ge, et al., 1994; Gutman, McLoyd, and Tokoyawa, 2005; see also Chapter 6) and the fact that the United States has the highest rate of child poverty among 25 economically devel- oped nations (United Nations Childrenâs Fund, 2001), policies that reduce poverty should have a particularly high priority. Numerous policies can reduce family poverty and its effects, either by increasing available income through such programs as the Earned Income Tax Credit, unemployment insurance, Temporary Assistance to Needy Families, or federal housing subsidies, or by helping address nutritional needs through such programs as food stamps, the School Lunch Program, and the Special Supplemental Nutrition Program for Women, Infants, and Children. Similarly, policies that could contribute to reducing health disparities and differential access to health care services (e.g., Medicaid, expansion of the State Childrenâs Health Insurance Program), as well as public health policies that minimize harmful environmental factors, such as exposure to neurotoxins (e.g., lead), warrant consideration. Policies that promote increased access to early child- hood education programs could set children, particularly impoverished chil- dren, on a positive life course. Finally, policies that support families, such as parental or family leave policies, access to quality child care, affordable transportation, recreational areas, and safe neighborhoods, facilitate sup- portive families and communities. Policies shifting schools and the juvenile justice system away from the use of punishment and toward the use of positive methods of developing desirable social behavior are also needed. For example, the Los Angeles Unified School District recently adopted a policy that requires the imple- mentation of systems of positive reinforcement in schools as an alternative to punishment (Los Angeles Unified School District, 2007). The policy is based in part on empirical evidence that punitive practices increase vandal- ism and antisocial behavior (Mayer, 1995). Expanding implementation of relevant public policies should be part of a national implementation strategy to support prevention of MEB disorders and promotion of mental health. At the same time, research on the impact
IMPLEMENTATION AND DISSEMINATION 325 of policies and on strategies for achieving effective public policies should be a portion of the nationâs prevention science portfolio. RESEARCH NEEDS Implementation and Dissemination Research An important step prior to implementation should be the availability of effectiveness studies. As more programs have shown efficacy in controlled trials, a next stage in prevention is studies of effectiveness under real-world conditions (Institute of Medicine, 1994). This research focus, often called âtype 2 translationalâ research, occurs after efficacy has been established and focuses on factors associated with the adoption and use of scientifically validated interventions by service systems (Green, 2007). It also includes consideration of maintenance and sustainability issues at the practice level that can be used to guide implementation. In the real world, translation of science-based practices often stumbles, largely unguided, toward uneven, incomplete, and disappointing outcomes. Translational research explores the factors that influence the Âquality of implementation; in such studies, implementation quality itself may be the outcome. A developing âscience of implementationâ (Dane and ÂSchneider, 1998; Durlak, 1998; Domitrovich and Greenberg, 2000) emphasizes the potential to advance the adoption of effective programs and Â redesign of health systems to ultimately improve health (Madon, Hofman, et al., 2007; C Â hambers, 2008). There is increasing knowledge regarding a variety of factors that influence implementation quality and recognition that better quality implementation leads to improved outcomes for children (Durlak, Weissberg, et al., 2007). Translational research related to school-based interventions should focus on a variety of factors: the decision-making process, the curriculum model or policy and the implementation support system, nonprogram fac- tors, such as characteristics of teachers and students, and policies and regulations of school and governmental bodies. For example, a recent c Â ommunity-based study highlighted the interactive influences of high-Âquality implementation by teachers and level of principal leadership in influencing aggressive behavior in elementary school-age children (Kam, Greenberg, and Wells, 2003). There are at least three conceptual models that may assist in guiding research questions, including those of the National Implementa- tion Research Network (Fixsen, Naoom, et al., 2005), the school ecologi- cal model (Greenberg, Domitrovich, et al., 2006), and the REACH model (Glasgow, Klesges, et al., 2004). In addition, Spoth and Redmond (2002) present a conceptual framework for scaling up preventive interventions and moving from effectiveness to
326 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Dissemination and Moving to Implementation Studies Scale Sustainability Adoption Effectiveness Studies Efficacy Studies Preintervention Fig11-1.eps FIGURE 11-1â Stages of research in prevention research cycle. implementation to achieve greater public health impact. They suggest three interrelated sets of research requirements and findings to accomplish popu- lation-based prevention: â(1) rigorously demonstrating intervention effec- tiveness; (2) attaining sufficient levels of intervention utilization in diverse general populations, requiring study of recruitment/retention strategies, cul- tural sensitivity, and economic viability; and (3) achieving implementation quality, involving investigation of adherence and dosage effects, along with theory-driven, intervention quality improvementâ (p. x). To accomplish this, it may be useful to view implementation as having three phases: adoption, sustainability, and moving to scale (see Figure 11-1). Prevention scientists, government organizations, state and community organizations, and com- munity leaders have major roles throughout this process. Ideally, the results of these phases will feed back to earlier areas of research. Specific research questions related to each of these phases warrant additional attention: â¢ Research questions related to the adoption of a prevention program into a service system, which routinely involves the formation of partnerships and the development of an infrastructure to support the technical, financial, administrative, monitoring, evaluative, and logistical needs related to the program.
IMPLEMENTATION AND DISSEMINATION 327 â¢ Research questions pertaining to sustaining the program once it is introduced in a service system. The ability to sustain a program relates to the organizational structures, practices, data monitoring, leadership, and related characteristics in place in the home institu- tion for the program. â¢ Research questions involving moving to scale, or understand- ing which steps facilitate the structures and funding necessary to expand the program to other sites. There are major challenges of introducing and taking effective pro- grams to scale, particularly in poor and underserved communities (Madon, Hofman, et al., 2007; Sanders and Haines, 2006), and clearly the current body of generalizable knowledge is inadequate to provide robust strategies for effective implementation across different populations, systems, and programs. Nevertheless, there is reason for optimism. First, the dearth of generalizable knowledge is a product of the lack of significant investment in scientific studies of the implementation and dis- semination process. Second, while the specific factors regarding successful program implementation may vary from case to case, there are many com- monalities in why organizations have difficulties adopting and sustaining prevention programs. For example, poor communities, minority popula- tions, and developing countries often lack professionally trained staff to deliver a program as originally designed, so successful implementation may need to identify unique program delivery agents using existing resources (Sanders and Haines, 2006). Also, there is general agreement in the field about shared dimensions of organizational change that are relevant across widely different interventions; these include system readiness for change, culture, and the role of leaders (Chambers, 2008). Third, the increasing use of more rigorous designs, such as randomized trials that test differ- ent implementation strategies (see Chapter 10), social network analysis ( Â Chambers, 2008), and the combined use of qualitative and quantitative data, is likely to lead to more precise implementation inferences around shared research questions. This information can be used as markers to guide the development of successful implementation efforts across diverse fields and settings. Implementation research turns the traditional efficacy/effectiveness research questions of prevention science into experimental questions about the process of implementation itself. To date the leading model for exam- ining implementation of prevention programs has been to focus on a sin- gle region, such as an urban school district or an entire state. With this approach, one can examine how process factors affect the adoption of a program over time, even if there are unique or novel factors operating in that particular system (Biglan, 2004). In addition, some randomized trial
328 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS designs are beginning to be used to study the process from effectiveness to large-scale implementation. For example, the community epidemiology model of Kellam and colleagues can examine questions related to effec- tiveness, sustainability, and moving to scale, along with randomization (Kellam, Koretz, and Moscicki, 1999; Kellam, 2000). This approach exam- ines variation in the community through the use of random assignment of the intervention conditions to different contexts and across time. To test effectiveness of a classroom-based intervention, for example, classes in a school can be randomized to intervention, and the impact for interven- tion and control classrooms can be compared in the first cohort (Brown, Wang, et al., 2008). This can be followed by examination of intervention sustainability by measuring the level of program fidelity that intervention teachers deliver in the second year with new students. Teachers remain in the same intervention condition, but the support structure in the schools for monitoring, superÂvision, and resource allocation changes from the first year to reflect the way such a program is likely to be delivered over time. Finally, scalability can be examined in a third cohort in which all the teach- ers implement the full intervention. Again, with resources allocated as one would anticipate in a scaled-up program, this third cohort can be used to compare the level of program fidelity as well as child outcomes with those of previous cohorts subjected to different levels of infrastructure support. This model, however, allows limited testing of the components of a higher level implementation strategy involving the full school district, since there is only one such district studied at a time (Kellam, 2000). Implementation Trials A valuable approach that would increase knowledge of successful imple- mentation strategies is to test alternative strategies using a randomized trial design (implementation trials). This would necessarily require multiple loca- tion and multilevel analyses to fully examine impact. One such implementa- tion trial is comparing the CTC model (see Box 11-1) to an implementation plan with passive assistance (Hawkins, 2006). In Project Adapt, Smith, Swisher, and colleagues (2004) tested two types of implementation of Life Skills Training (LST) using group-based random assignment. They compared a standard implementation model in which the LST curriculum stood apart from the day-to-day teaching activities with an infusion model that integrated the curriculum into traditional courses. Three rural school districts were randomly assigned to the traditional LST condi- tion, the infused LST condition, and a control condition. There was some suggestion of beneficial impact of both intervention conditions against con- trol for girls in the first year of the study, although these findings generally disappeared by the second year and did not show at all for boys. There were
IMPLEMENTATION AND DISSEMINATION 329 few differences between the two intervention conditions, although this could have been due to the low statistical power for this school-based design. An ongoing randomized trial of two different methods of implementing an evidence-based program for foster care in California counties may shed new light on implementation approaches. The trial was driven by a Cali- fornia mandate to use evidence-based practices and interest in identifying ways to facilitate statewide implementation. Although this particular trial involves multidimensional treatment foster care (Chamberlain, Saldana, et al., in press), an evidence-based program that targets high-need children who are in state custody, it can also be enlightening for the implementa- tion of evidence-based prevention. Training had earlier been offered to all California counties, but only about 10 percent of the counties became early adopters, not unlike that of most novel interventions (Rogers, 1995; Valente, 1996). All the remaining 40 eligible counties were randomly assigned to one of two methods for implementation: a standard model and a community development team model, which used cross-county peer-to-peer support to address the administrative, financial, and logistical challenges in implement- ing the program. The evaluation is assessing whether the rate and length of time for adoption and sustainability is reduced by the team model, taking into account the dependence between team members. Research on Increasing Rates of Intervention Adoption and Participation The rate of adoption of a particular program across different commu- nities and the rate of participation in a community are major issues that affect levels of program implementation. A variety of potential approaches to increase these rates could be evaluated in future research. Encouragement Designs The general class of randomized âencouragement designsâ are ones that randomize individuals to different modalities of recruitment, incen- tives, or persuasion messages to influence their choice to participate in one or another intervention condition. Such incentives as cash or child care dis- counts have been used to encourage participation. An important advantage of these designs is that they allow one to take into account self-selection factors in examining impact (Yau and Little, 2001; Frangakis and Rubin, 1999; Barnard, Frangakis, et al., 2002). They also address whether targeted efforts to increase participation reach those most at risk (Brown and Liao, 1999). Randomized encouragement trials have been used to evaluate early versus late enrollment in Early Head Start (Administration for Â Children and Families, 2005), whether antiviral medications for HIV should adhere to a rigid regimen or be more flexible (HIV SMART AntiRetroviral Trial),
330 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS BOX 11-4 The Internet as a Potential Tool for Wide-Scale Dissemination of Preventive Interventions The enormity of need for mental health services often produces a type of paralysis: since it is not feasible to train enough providers to treat all individuals with mental, emotional, and behavioral disorders, how can preventive interven- tions be provided to those at risk? This dilemma is caused in part by the exclusive reliance on consumable interventions, such as face-to-face services, and the use of medications. Once a prevention or therapy session is over, no other individual can benefit from that hour of contact. Once a medication is consumed, no one else can benefit from its therapeutic effect. The development and implementa- tion of interventions delivered via the Internet offers the promise of an approach to make interventions available on a continual basis to a wide range of young people at minimal cost while addressing several dissemination and implementa- tion challenges. Fidelity: The fidelity of Internet interventions is inherent as the material on the computer screen remains the same, no matter how many times it is used. The content of the intervention can be shared widely exactly as tested in randomized control trials. Scalability: An Internet intervention can be shared with literally thousands of users beyond the locality in which it was created, while remaining accessible to the original locality. The site of a proven Internet intervention can be immediately opened to use by anyone with web access, which also allows effectiveness evalu- ation on a wide scale. Sustainability: The cost of maintaining a website hosting an evidence-based preventive intervention is relatively modest, especially if the site is an automated, self-help intervention. and whether strategic, structural engagement of adolescents increases completion of family therapy more than traditional engagement methods ( Â Szapocznik, Perez-Vidal, and Brickman, 1988). Such designs may have value in exploring ways to increase the reach of prevention programs. In the years since publication of Reducing Risks for Mental DisÂorders: Frontiers for Preventive Intervention Research (Institute of Medicine, 1994), a modest number of experimental tests have aimed at increasing individual- or Âfamily-level participation rates for a preventive intervention. For example, motivational interviewing techniques have been used in trials in an attempt to engage parents around problems or issues that they can relate to their own children (Dishion, Kavanagh, et al., 2002).
IMPLEMENTATION AND DISSEMINATION 331 Accessibility: Internet interventions can simultaneously serve users across a community, a state, the nation, or the world, at any time of the day or night, includ- ing holidays and weekends. Stigma: The availability of Internet interventions that are used in the privacy of oneâs own home, educational or work setting, or using a public access computer makes these interventions more likely to be used by people who would not come to a mental healthâoriented program. Reaching multicultural, multilingual communities: Internet interventions can be implemented relatively easily in multiple languages. Similarly, advances in technology now make it possible to create Internet interventions that require a minimum level of reading or writing. The use of video, graphics, and audio allow the creation of Internet interventions that can be used by individuals at any edu- cation level. Internet interventions also have limitations. One of the most troublesome is the lack of access to the web by many low-income, low-education groups. However, Internet access is increasingly available via mobile devices, such as cell phones. Many developing countries have skipped the stage of land-line phones and moved directly to cell phones. As is the case for other venues, Internet interventions will not be effective in preventing all types of MEB disorders. It is useful to think in terms of âmarket segmentation,â in which specific means of reaching populations at risk will need to be developed and evaluated to see which is most effective for which population. Nevertheless, to help make prevention feasible, one must think beyond traditional interventions and harness the power of advanced communica- tion media, such as the Internet. Use of Current Technologies The advent of the Internet and modern use of technology presents new opportunities for both dissemination and research. Broadcasting the avail- ability of accessible web-based or CD-ROM programs or making imple- mentation resources (e.g., training, manuals) available could potentially increase the use of prevention programs. Implementation of interventions via the web has the potential to address several implementation barriers (see Box 11-4). Because online interventions can occur anonymously, these technologies also have the potential to be less stigmatizing, a significant potential barrier to participation.
332 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Identification of Early Adopters Rogers (1995) identified general factors that affect or influence the diffusion of innovations. This early work on program diffusion was based on a synthesis of careful observation from case studies. One major finding is that early adopters share traits that can be readily measured or inferred from behaviors or attitudes. While this earlier work was observational in nature and did not attempt to influence adoption itself, assessments can be used to identify communities, organizations, institutions, families, as well as individuals who are most likely to be early adopters of such programs. Thus identification of those likely to be early adopters and targeting preven- tion efforts to these groups represent a potential strategy to affect program adoption. Use of Opinion Leaders More recently, the same principles underlying research on diffusion of interventions and social influence have been used proactively to increase the adoption of prevention programs and test adoption strategies in group-based randomized trials. One approach used early in HIV prevention is to target opinion leaders in a community who would themselves deliver peer-to-peer messages to promote increased program adoption. Kelly, St. ÂLawrence, and colleagues (1991), for example, successfully identified and then trained gay opinion leaders in rural communities to encourage safe sexual practices. These leaders were able to modify HIV risk behaviors in their communi- ties. Also, media campaigns for HIV prevention in developing countries are using soap operas in which leading actors talk openly about the use of condoms and getting tested for HIV (Valente, 1996). A similar approach is now being used in approaches to youth suicide prevention; teenage leaders are trained to deliver messages to both peers and adults in their community aimed at increasing help seeking among sui- cidal youth. Suicidal youth are often much less likely to talk to adults than are nonsuicidal youth (Wyman, Brown, et al., 2008), yet the vast majority of youth tell a friend before committing suicide. A general strategy for reducing suicide is to increase willingness to talk to a trusted adult by both suicidal youth and their friends. One such program (Sources of Strength) is designed to change peer norms about secrecy and disclosure surrounding distressed youth. A first implementation step is to identify peer leaders from diverse social networks. The program then modifies norms by having each of the peer leaders identify trusted adults in their own lives to whom they would turn at times of stress.
IMPLEMENTATION AND DISSEMINATION 333 Market Research Many evidence-based prevention programs are delivered to small por- tions of the population. A small number of state agencies, schools, com- munities, or families select programs with the highest levels of evidence, opting instead for programs that have less evidence, or no program at all. One promising approach to improve program reach to individual families is to integrate business models into prevention to address consumer needs from the beginning (Rotheram-Borus and Duan, 2003). By following a prevention service development model that integrates consumer preferences from the beginning (Sandler, Ostrom, et al., 2005), the research team can aim for effectiveness and large-scale implementation from the start of the product development cycle. Similarly, there is a need for greater consideration of the most effective metrics to report outcomes to the public. Although effect size may be the most appropriate metric for studies of indicated interventions in which all participants begin with a substantial rate of symptoms, it may be a poor m Â etric for universal interventions. In universal interventions, it is usually the case that a large percentage of the population begins with low levels of symptoms, and thus it is unlikely (at least in the short term) that much of this population will benefit from the intervention. In most cases it is only in the higher symptom group of the population that larger effect sizes will be obtained (Wilson and Lipsey, 2007). Thus, for universal interventions, alternative methods are needed to convey the practical and social policy sig- nificance (Davis, ÂMacKinnon, et al., 2003; McCartney and Rosenthal, 2000). Cost-effectiveness is one such metric, as universal interventions may achieve more benefit in relation to their cost given their large reach. Naturalistic Large-Scale Public Health Research Although their internal validity makes them valuable science, random- ized control trials do not always have good external validity. Furthermore, much academic research is rarely applied to the day-to-day world. Science can often benefit from the experience of everyday clinical observations. For example, in 1982 when clinical observations in a community mental health setting found an extraordinary number of children exposed to violence, a plethora of scientific research projects confirmed this observation, culminat- ing in several large-scale strategies to prevent these children from develop- ing mental health sequelae (Jenkins and Bell, 1997; Bell, 2004). In addition, communities often implement programs because they are based on extensive clinical wisdom and have widespread community sup- port. Research designed to empirically test programs being implemented in naturalistic environments could identify approaches that are readily imple-
334 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS mentable by other communities. Gibbons, Hur, and colleagues (2007) have developed statistical methodology that provides some evidence in support of such interventions. Conversely, randomized control prevention trials can also inform public health practice. For example, a violence prevention trial, Aban Aya (Flay, Graumlich, et al., 2004) informed a Chicago public school violence prevention initiative with teenage mothers (Bell, Gamm, et al., 2001), which demonstrated significant reductions in pregnant teenage dropout rates. In addition, most teens had only one child despite becoming a mother at very young ages (Lamberg, 2003). CONCLUSIONS AND RECOMMENDATIONS There have been clear advances in implementing effective programs since the publication of Reducing Risks for Mental Disorders: Frontiers for Preventive Intervention Research (Institute of Medicine, 1994). Indeed, the knowledge base on effective prevention programs at that time was very thin. However, the levels of effective implementation are much lower than the availability of tested interventions suggests. Conclusion: Implementation of effective preventive interventions is hampered by lack of ongoing resources and competing priorities of the service systems or communities that could implement them. One of several contributors to the relative lack of implementation is lack of empirical evidence regarding how to effectively approach imple- mentation. A critical next phase of research needs to examine methods for enhancing the implementation of effective programs. The prevention research cycle proposed in the 1994 IOM report assumes a âhierarchical scientist-as-expert perspective and portrays scientists as separate agents conducting research on âsubjectsâ and âgroupsââ (Dumka, Mauricio, and Gonzales, 2007). Although the stages of research in the model require the cooperation of individuals and organizations, the model did not specifically address the relationships and collaborative processes that are critical to accomplishing each stage (Dumka, Mauricio, and Gonzales, 2007). For implementation to be successful, there needs to be strategic input from science, policy, and practice perspectives that builds on the scien- tific knowledge base. Evidence is needed on how to make implementa- tion occur in communities, the policy directives that promote or enforce the use of Â evidence-based programs and data systems, and the effective adoption and sustainability of programs in practice (Greenberg, 2004). Important progress has been made, and there are now new opportunities to make partnerships between scientists, policy makers, and practitioner
IMPLEMENTATION AND DISSEMINATION 335 communities to transport effective prevention programs into community settings. Additional research is needed to identify core components shared across programs. Major implementation challenges suggest new avenues of research. Conclusion: Knowledge about effective strategies for implementing or adopting evidence-based prevention interventions is limited. New approaches to implementation represent the frontier of prevention research. Recommendation 11-1: Research funders should support experimen- tal research and evaluation on (1) dissemination strategies designed to identify effective approaches to implementation of evidence-based programs, (2) the effectiveness of programs when implemented by communities, and (3) identification of core elements of evidence-based programs, dissemination, and institutionalization strategies that might facilitate implementation. Knowledge gained from evaluation of implementation approaches will be more generalizable if it is conducted in multiple settings. A number of evidence-based interventions are viable candidates for implementation. Evaluations that involve partnerships between states or communities ready to implement interventions and researchers could yield valuable results. Recommendation 11-2: Research funders should fund research on state- or community-wide implementation of interventions to promote mental, emotional, or behavioral health or prevent MEB disorders that meet established scientific standards of effectiveness. Although there are many evidence-based models, it is not clear how gen- eralizable they are to groups other than the ones with which they were tested. Interest in an intervention is likely to be greater if it is culturally relevant and embraced by the community. Lack of relevance may contribute to interven- tions being implemented with limited fidelity and resultant limited outcomes. Addressing this may include replication with new populations as well as examining versions that strengthen the cultural competency of interventions. Conclusion: Despite multiple dissemination venues, evidence-based interventions have not been implemented on a wide-scale basis. Where interventions have been implemented, they are often not implemented with fidelity, with cultural sensitivity, or in settings that have the c Â apacity to sustain the effort.
336 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Conclusion: Little research has addressed the question of how trans- portable evidence-based interventions developed for one ethnic group are to a range of ethnic and cultural groups. Recommendation 11-3: Research funders should prioritize the evalua- tion and implementation of programs to promote mental, emotional, or behavioral health or prevent MEB disorders in ethnic minority com- munities. Priorities should include the testing of culturally appropriate adaptations of evidence-based interventions developed in one culture to determine if they work in other cultures and encouragement of their adoption when they do. Finally, multiple opportunities for naturalistic research could enrich the prevention portfolio and convincing evidence that collaborations between researchers and communities can increase the relevance and sustainability of interventions, including through efforts to adapt existing evidence-based interventions. Recommendation 11-4: Researchers and community organizations should form partnerships to develop evaluations of (1) adaptation of existing interventions in response to community-specific cultural characteristics; (2) preventive interventions designed based on research principles in response to community concerns; and (3) preventive interventions that have been developed in that community, have dem- onstrated feasibility of implementation and acceptability in the com- munity, but lack experimental evidence of effectiveness. On a practical level, for tested preventive interventions to become widespread, the available research suggests that successful interventions should include at least the availability of published material, such as hand- books, curriculum, and manuals describing the intervention and prescribing actions to be taken; certification of trainers or an electronic training sys- tem; high-quality, data-driven technical assistance; implementation fidelity measures; dissemination efforts that are organized around marketing and delivery; an information management system; and community demand for systems that work. In addition to development and implementation of effective programs, the nation needs to support implementation of policies and broad preven- tion principles in order to create a comprehensive, sustained approach to prevention. Policies that support low-income families and promote healthy development are needed as the basic foundation for such an approach.