FUTURE DIRECTIONS
Future Directions for Dissemination and Implementation Science: Aligning Ecological Theory and Public Health to Close the Research to Practice Gap
Marc S. Atkins, Dana Rusch, and Tara G. Mehta
Institute for Juvenile Research, University of Illinois at Chicago
Davielle Lakind
Department of Psychology, University of Illinois at Chicago
Dissemination and implementation science (DI) has evolved as a major research model for childrenâs mental health in response to a long-standing call to integrate science and practice and bridge the elusive research to practice gap. However, to address the complex and urgent needs of the most vulnerable children and families, future directions for DI require a new alignment of ecological theory and public health to provide effective, sustainable, and accessible mental health services. We present core principles of ecological theory to emphasize how contextual factors impact behavior and allow for the reciprocal impact individuals have on the settings they occupy, and an alignment of these principles with a public health model to ensure that services span the prevention to intervention continuum. We provide exemplars from our ongoing work in urban schools and a new direction for research to address the mental health needs of immigrant Latino families. Through these examples we illustrate how DI can expand its reach by embedding within natural settings to build on local capacity and indigenous resources, incorporating the local knowledge necessary to more substantively address long-standing mental health disparities. This paradigm shift for DI, away from an overemphasis on promoting program adoption, calls for fitting interventions within settings that matter most to childrenâs healthy development and for utilizing and strengthening available community resources. In this way, we can meet the challenge of addressing our nationâs mental health burden by supporting the needs and values of families and communities within their own unique social ecologies.
We were making little progress explaining to a classroom teacher why we were not seeing one of her students in our hospitalâs outpatient clinic. The child was 7 years old and referred to us because of his aggressive behavior at school. In addition to work with his parent, we proposed a behavior management plan for his school. But his teacher wanted none of it, and now the school principal asked to meet with us. The principal listened carefully as we explained our reasoning: âThere is little we can do in a clinic that will improve his behavior at school.â She nodded and said she understood. âYou want to work with my teachers,â she said, more statement than question. âWe have a lot of fighting at recess. Can you fix our playground?â
This incidental conversation more than 20 years ago launched a program of research that informs this commentary. Our response to this principalâs question was a qualified yes. Yes, we knew how to reduce aggression through improved supervision and engaging activities. But could we reduce aggression in this urban school relying only on the schoolâs own resources? Although there was scant literature to guide that question, we were hardly alone in our interest to bridge the research-to-practice gap. Inspired by the groundbreaking studies of John Weisz and colleagues, who first reported the lack of clinical representation (i.e., sample, provider, setting) in research studies in clinical child psychology (Weisz, Jensen-Doss, & Hawley, 1995), we were challenged not only to show positive outcomes but also to increase access and to sustain the services in these critically underserved communities (Atkins, Gracyzk, Frazier, & Abdul-Adil, 2003).
Subsumed under the rubric of dissemination and implementation science (DI; Schoenwald, McHugh, & Barlow, 2012), there is an emerging consensus for a more rapid integration of effective practices into usual care to close the research to practice gap and to reduce the nationâs long-standing mental health burden (Glasgow & Chambers, 2012; Kazdin & Blase, 2011). To date, DI has relied heavily on social diffusion theory (to denote that people and their social networks matter) and organizational theory (to denote that settings matter); however, we argue that the overfocus on promoting packaged programs in the form of evidence-based programs (EBPs) does not successfully integrate the knowledge of settings and persons toward maximal impact. In this article, we describe a process that aligns ecological theory with a public health model to address long-standing mental health disparities. We describe how this process evolved from our work in urban schools, and we close with a description of a planned program of research to support Latino immigrant origin families. Our examples expand DI research from an emphasis on how to bring this program to that setting toward an ecologically driven science that prioritizes the needs and resources of settings that matter most to youth and families.
DI RESEARCH TO BRIDGE THE RESEARCH-TO-PRACTICE GAP
More than two decades ago, a series of national reports documented the slow pace of advances in mental health research and the need for a new paradigm to promote an integration of science and practice. Crystalized in the first-ever Surgeon Generalâs report on mental health, a groundbreaking review of the pervasive and unrelenting mental health burden facing our nation (U.S. Department of Health and Human Services, 1999), these issues were further articulated in two reports sponsored by the National Institute of Mental Health (NIMH) National Advisory Mental Health Council to broaden the NIMHâs research portfolio to promote public health (National Institutes of Health, 1999) and to align the differing cultures of science and practice (National Institutes of Health, 2000). Recent NIMH strategic plans also promote a public health approach to enhance the use of innovative prevention and intervention programs for those most in need and strengthen the public health impact of research (Insel, 2009; NIMH, 2015).
Bridging science and service, however, has proved more elusive than originally expected. Often referred to as a 17-year saga from âbench to bedsideâ (Wolff, 2008), the attempt to bridge science and practice has been described as a âvalley of deathâ (Meslin, Blasimme, & Cambon-Thomsen, 2013), reflecting the few successful enduring outcomes. Green (2008) proposed a new set of priorities for research that incorporates practitioner perspectives throughout the research pipeline (Green, Ottoson, Garcia, & Hiatt, 2009); remarkably similar to a proposal to advance researcherâpractitioner collaboration in clinical child psychology (Weisz, Chu, & Polo, 2004). More broadly, DI has become a prominent model in mental health research to examine potential mechanisms that can effectively and efficiently bridge the research-to-practice gap and lead to the integration of EBPs into routine mental health services. Multiple factors have been identified that impact the dissemination and implementation of effective services into community settings, including individual characteristics (e.g., provider attitude; Aarons, 2004), organizational characteristics (e.g., social context; Glisson, 2002), and public policy characteristics (e.g., financial resources; Massatti, Sweeney, Panzano, & Roth, 2008). Two research models have been especially influential in DI and are likely to remain prominent in future research: Social diffusion theory and organizational theory.
DI RESEARCH: THE INFLUENCE OF PERSONS AND SETTINGS
Diffusion of innovation theory (Rogers, 2003) and social network theory (Wasserman & Faust, 1994) describe the process by which new practices are adopted by an individual within a social setting. Empirically validated across diverse settings (e.g., agriculture, medicine, education), these theories have guided research demonstrating that information (or innovation) is adopted and spread among individuals in a predictable pattern and is influenced by individuals in particular roles within a network of relationships (Burt, 1999). Social networks support the adoption and use of innovative practices by providing a context within which individuals make decisions regarding use of practices. These decisions are based in part on peer influences leading to a âtipping pointâ of use, after which the former innovation becomes standard practice (Berwick, 2003; Green et al., 2009; Neal, Neal, Atkins, Henry, & Frazier, 2011).
Concurrent with research on diffusion theory and social network processes, organizational theory has emphasized the importance of setting characteristics for the uptake and implementation of innovative practices (see Aarons, Hurlburt, & Horwitz, 2011; Damschroder et al., 2009). Glisson (2002) developed a widely used theory of organizational context that has been empirically validated in several contexts (e.g., childrenâs social services, emergency rooms) and associated with more positive client outcomes, positive work environments, lower staff turnover, and higher quality of services (e.g., Glisson & Green, 2011; Glisson, Hemmelgarn, Green, & Williams, 2013; Glisson et al., 2008). In a recent study, an organizational intervention derived from this work (Availability, Responsiveness, Continuity; Glisson & Schoenwald, 2005) was shown to enhance the effectiveness of Multisystemic Therapy for seriously conducted disordered youth (MST; Henggeler, Schoenwald, Borduin, Rowland, & Cunningham, 2009), in high-poverty rural communities (Glisson et al., 2010). Specifically, therapist fidelity to MST was similar across conditions; however, youth in counties that implemented MST plus Availability, Responsiveness, Continuity improved at a faster rate in the first 6 months of treatment, and were significantly less likely to enter an out-of-home placement, than youth in counties that implemented only MST.
MISGUIDED ASSUMPTIONS OF DI IN MENTAL HEALTH SERVICES RESEARCH
As exciting as these research findings are for efforts to bridge science and practice, there are several limitations in the application of these literatures to DI that compromise the public health impact of research. The first is the assumption that EBPs are the gold standard of intervention as indicated by the predominant emphasis on identifying and targeting factors influencing EBP program adoption. There are two problems with this assumption. First, it does not acknowledge that research on which these programs were based has largely failed to accommodate the realities of community practice (Weisz, Jensen-Doss, & Hawley, 2006). Thus, rather than attributing failure to adopt an EBP to deficits in practice sites (e.g., low readiness for change; Weiner, Amick, & Lee, 2008), the most parsimonious explanation for poorer outcomes of EBP implementation in usual care is offered by Weisz and colleagues as a failure of research-derived practices to match the characteristics of usual care (Weisz, Ugueto, Cheron, & Herren, 2013). This is not to say that EBPs have no relevance to mental health practice. As Weisz and colleagues have shown, EBPs can be aligned with usual care when the values and expertise of providers are recognized and supported and the context of usual care settings are accommodated (Santucci, Thomassin, Petrovic, & Weisz, 2015; Weisz, Krumholz, Santucci, Thomassin, & Ng, 2015). However, context-specific services are the exception not the rule in DI research, which suggests the need for a large dose of humility to acknowledge that âwhat works for whom under which conditionsâ (Paul, 1967) has been a long-standing inspirational goal that is largely unad-dressed by current research (Hoagwood, Atkins, & Ialongo, 2013; Weisz, Ng, & Bearman, 2014; Weisz et al., 2013).
Second, with more than 200 interventions for children and adolescents currently listed in SAMHSAâs National Registry of Evidence-Based Programs and Practices (www.nrepp.samhsa.gov), it is unclear how community providers would select the most relevant intervention to meet the needs of their clients. Moreover, even if providers were able to implement multiple EBPs with perfect fidelity, there would be many children with problems that extant EBPs do not address. This was demonstrated powerfully by Chorpita, Bernstein, and Daleiden (2011), who showed that 86% of youth in a statewide clinical sample were not coverable when matched on presenting problem, age, gender, and ethnicity, as culled from 435 randomized clinical trials of EBPs. Over the past decade, Chorpita and colleagues have made considerable strides to broaden the reach of EBPs through a distillation of best practices for use by community practitioners (see Becker et al., 2015; Chorpita & Daleiden, 2009). As we illustrate, to complement this effort, an ecological framework could provide situational specificity to match best practices with setting-specific parameters.
A third misconception (to the extent that it is considered at all) is that wide-scale implementation of EBPs will effectively reduce the population prevalence of mental health issues. Dodge (2009) described the challenges of bringing interventions to scale, observing, for example, that the Fast Track preventive intervention (Conduct Problems Prevention Research Group, 2011) would potentially reduce the population prevalence of conduct disorder by just 2% in spite of its individual-level treatment record of reducing conduct disorder by 50%. This is due to challenges identifying at-risk children, serving nontreatment seekers and non-treatment completers, and family mobility. He also illustrated the unexpected consequences of scaling up interventions by describing a statewide push to reduce class sizes that resulted in a shortage of adequately credentialed teachers (Bohrnstedt & Stecher, 1999).
Given that our own work focuses on children and families in particularly vulnerable circumstances, we are especially aware of the limitations of DI for under-served populations. For example, selecting appropriate EBPs becomes more challenging in community mental health practice in high-poverty communities where presenting problems are complex and comorbidity is the norm (Bradley & Corwyn, 2002; Xue, Leventhal, Brooks-Gunn, & Earl, 2005). Furthermore, although children and families in need of mental health services from all communities face significant barriers to seeking and receiving services in usual care settings (Kataoka, Zhang, & Wells, 2002), these barriers are more complicated and entrenched for children and families in poverty (Harrison, McKay, & Bannon, 2004). As currently conceptualized, the power and viability of DI will remain limited in its ability to address the public health needs of the majority of children and families, especially those who are already most vulnerable.
ECOLOGICAL THEORY: EMBRACING PERSONS AND SETTINGS
One way to overcome the limitations of DI research is to align interventions with settings important to children and families. This would lead ultimately to an understanding of setting-specific and person-specific factors associated with adoption and sustainability of practices across the broad array of settings and situations that influence childrenâs mental health. Ecological theory describes the dynamic interrelations among personal and environmental influencesâthe social ecologyâand the direct effects of persons and settings on childrenâs health and mental health. The ecological framework we propose for DI draws on concepts from Bronfen-brenner (1977, 1979) and Kelly (1966, 2006), and from the biological study of sustainability (Dodson et al., 1998).
The now-classic image of Bronfenbrennerâs concentric rings illustrates that individualsâ behaviors are influenced by the settings in which they are embeddedâtheir families, schools, neighborhoods, nationsâand that we cannot fully understand these behaviors without accounting for the additive and interactive effects of these contextual factors. Kelly emphasized the impact of settingsâ physical and sociocultural characteristics on individuals, as well as the reciprocal impact individuals have on the settings they occupy. Indi...