DISSEMINATION OF EVIDENCE-BASED INTERVENTIONS THROUGH COMMUNITY-UNIVERSITY PARTNERSHIPS: EARLY FINDINGS FROM THE PROSPER PREVENTION TRIAL
Chair: Richard Spoth
DISSEMINATION OF EVIDENCE-BASED INTERVENTIONS THROUGH COMMUNITY-UNIVERSITY PARTNERSHIPS: EARLY FINDINGS FROM THE PROSPER PREVENTION TRIAL. Richard Spoth1, J. David Hawkins2, 1Partnerships in Prevention Science Institute at Iowa State University, Ames, IA United States; 2Social Development Research Group at University of Washington, Seattle, WA United States
Many of the youth and family-focused interventions currently utilized in U.S. communities lack demonstrated empirical support (Hallfors & Godette, 2002). Moreover, there is limited research on effective strategies and supportive infrastructures for quality implementation of evidence-based interventions (EBIs, see Wandersman et al., 1998). Randomized, controlled studies of community-based partnerships focused on the diffusion of youth and family-focused EBIs are very rare (Spoth & Greenberg, 2004). The studies presented in this symposium address these gaps in the relevant research.
This symposium describes a large-scale prevention effectiveness trial evaluating community-university partnership-based implementation of EBIs. PROSPER (PROmoting School/community-university Partnerships to Enhance Resilience) partnerships are structured with teams at three levels: (1) small, strategic community teams that implement family-focused and school-based EBIs selected from a menu, using a local “linking agent” approach, (2) university research teams, and (3) Prevention Coordinators, who provide technical assistance to community teams and serve as liaisons between community teams and university researchers. The PROSPER randomized study includes a sample of 28 communities in Iowa and Pennsylvania; it addresses two primary aims: (1) to evaluate the effectiveness of partnerships in producing positive youth and family outcomes through EBIs, and (2) to learn what factors influence community team effectiveness (e.g., task orientation), the quality of EBI implementation, and the sustainability of quality implementation. To address the first aim, data are being sequentially collected from (a) two cohorts of students in project school districts, (N 13,000) and (b) a randomly selected subsample of families from the second of those cohorts. To address the second aim, information is gathered at multiple time points in intervention communities, from local PROSPER team members, direct supervisors of the local team members, Prevention Coordinators, and schools (primarily focusing on their prevention policies and resources).
Following a summary of the linking-agent conceptual framework for community-university partnership implementation of EBIs, the three symposium papers will summarize findings to date concerning three major questions:
•What predicts optimal functioning in the small, strategic community teams?
•What is the quality of community team-led implementation of the family-focused and school-based EBIs, as indicated on a range of measures?
•What are the effects of the family-focused EBI on proximally targeted outcomes?
PREDICTING LOCAL PREVENTION TEAM FUNCTIONING: LONGITUDINAL ANALYSES OF PROSPER. Sarah Meyer-Chilenski1, Mark Feinberg2, Mark Greenberg2, 1Pennsylvania State University, University Park, PA United States; 2Pennsylvania State University, State College, PA United States
This paper addresses the symposium´s first question concerning predictors of optimal functioning in PROSPER´s small, strategic community teams. Specifically, it describes how characteristics of individual members, teams, and communities at the stage of team formation predict how teams are functioning six months later. Although numerous studies have cross-sectionally examined factors that influence how collaborative teams operate, there has been very little longitudinal data regarding collaborative processes. We hypothesized that community readiness, as well as team member attitudes towards prevention, skills, and experience in collaborations would predict the quality of team functioning.
The data comes from over 150 interviews with 14 local community prevention team members. Members were interviewed within one month of team formation and 6 months later, when teams were planning the implementation of family-focused EBIs. Reliable scales from prior research were used at Wave 1 to measure community readiness, attitudes towards the value of prevention, expectations for PROSPER efficacy in the community, skills useful in collaborative community work, and experience in similar community partnerships. Census and school level data were used to assess demographic characteristics. At wave 2, interviews focused on multiple domains of prevention team functioning: the team´s ability to set a clear organizational structure, the level of cohesion and positive work relations among team members, and the team leaders´ ability to engage team members, and foster productive collaboration.
Hierarchical regression models were utilized with team as the unit of analysis, in which poverty and education levels of the community were entered in Step 1, followed by perceived community readiness, prior experience with collaboration, attitudes toward prevention, and skills. Results indicated that community poverty accounted for substantial variance (35% to 50%) of variance in team leadership and cohesive task orientation, but not team goals. Team members´ attitudes toward prevention also accounted for substantial variance of all outcomes (avg. R2 = .30).
To further explore the effect of poverty, we examined associations with a range of variables relating to other community, school, and team factors, but were unable to discover potentially mediating or proxy variables. Our discussion will focus on (1) considering how poverty might affect the dynamics of community team functioning, and suggest how future research might examine this issue further, and (2) the impact of initial attitudes towards prevention and its relevance to team member selection at the formation stage of partnerships.
MULTIPLE INDICATOR ASSESSMENT OF QUALITY IMPLEMENTATION OF EVIDENCE-BASED INTERVENTIONS BY COMMUNITY TEAMS. Richard Spoth1, Scott Clair1, Max Guyll1, Catherine Lillehoj1, Cleve Redmond1, 1Partnerships in Prevention Science Institute at Iowa State University, Ames, IA United States
The second question addressed in this symposium concerns community team outcomes on indicators of the quality of implementation of family-focused and school-based evidence-based interventions (EBIs). Successful diffusion of EBIs requires high quality implementation across communities. The current study evaluated implementation quality in the delivery of the EBIs selected by community teams from a menu, including: the Strengthening Families Program: For Parents and Youth 10-14 (SFP 10-14) and three school-based interventions, Life Skills Training, Project Alert, and All Stars.
In the case of high quality implementation with universal family-focused EBIs, the first implementation challenge is recruitment of general population families. For Cohort 1, PROSPER community teams recruited 1,387 family members, or 17.5% of all families of 6th graders. This recruitment rate is conservatively estimated, calculated with a count of all families of 6th graders in the community—whether or not they were, technically, eligible. It lies at the high end of the range for multi-session interventions implemented by researchers, and is higher than the typical level of community team-based recruitment for health promotion programming reported in rigorous studies (e.g., Saunders et al, 2003). As expected, technical assistance from Prevention Coordinators and the quality of the recruitment were most predictive of recruitment rates.
Results also showed that implementation quality was uniformly high, with SFP 10-14 facilitators covering 91% of program material overall; the three separate program segments focusing on youth, parents, and families had adherence rates of 91%, 90%, and 93%, respectively. Implementation adherence rates also were high for the three school-based EBIs, averaging 87%.
Preliminary analyses revealed no significant relationships between implementation adherence and factors that ordinarily might negatively affect implementation of the EBIs, suggesting the positive influence of the PROSPER partnership model linking university researchers, Extension personnel, and communities. Specifically, no significant relationships emerged between adherence and a wide range of variables pertaining to community SES, community team factors, and family factors.
Finally, recently-gathered evidence indicates the teams´ are sustaining implementation from local funding. That is, most teams have secured funds for continued implementation of the family-focused EBI midway through the five-year project.
Taken together, these preliminary data support the feasibility and implementation-related effectiveness of the PROSPER partnership-based diffusion model for universal EBI delivery.
PROXIMAL OUTCOMES OF A UNIVERSAL, FAMILY-FOCUSED EVIDENCE-BASED INTERVENTION IMPLEMENTED BY COMMUNITY TEAMS. Cleve Redmond1, Richard Spoth1, Chungyeol Shin1, Scott Clair1, Yan Guo1, 1Partnerships in Prevention Science Institute at Iowa State University, Ames, IA United States
This paper addresses the symposium´s third question regarding the proximal outcomes of the family-focused evidence-based intervention (EBI). The youth and family EBIs were implemented sequentially. In the first year of the project, the local teams selected a universal family-focused intervention for families of 6th graders from a menu of three universal EBIs. In the second year, teams implemented family-focused booster sessions and a school-based EBI (Grade 7). This paper examines the proximal outcomes targeted by the family-focused intervention, based on in-school assessments with the first cohort of students.
Participation rates for the in-school assessment were 89% in Wave 1 (N=5,721) and 88% in Wave 3 (N=5,667). The Wave 3 data collection took place shortly after the implementation of the family-focused boosters during the Spring of Grade 7. Data were collected in one class period at each data collection point.
Ten proximal outcomes targeted by the family-focused EBI were assessed across four domains: intervention-targeted parenting behaviors (problem solving and parent involvement in child activities); child management practices (general child management, consistent discipline, harsh discipline, and child monitoring); parent-child affective quality (affective quality with mother, affective quality with father, and affective quality with parents); and family cohesion. To assess the proximal outcomes, multi-level ANCOVA models were conducted. All intervention-control differences were in the expected direction; intervention effects on problem solving, general child management, consistent discipline, harsh discipline, and family-cohesion were significant at the .05 level. A preliminary dosage analysis compared those students in the intervention schools that attended at least one session of the intervention with all of the students in the control schools. This analysis showed the effect sizes for all measures were greater in the subsample participating in the intervention, as compared to intervention-control differences using the total sample. The reduced N and statistical power, however, decreased the number of statistically significant intervention-control differences in the dosage analyses to one (the parent involvement outcome). It is, however, noteworthy that effects on all three affective quality measures were marginally significant (p < .1) in these dosage analyses.
The high implementation quality achieved by the community teams likely contributed to the observed positive trends and beneficial effects on the proximal outcomes. The dosage analyses suggest that higher recruitment rates for the family-focused EBI and its booster sessions might have produced stronger effects on the proximally-targeted intervention outcomes.
CONCURRENT 2, INTEGRATING BIOLOGICAL AND SOCIAL FACTORS IN PREVENTION RESEARCH, Organized symposia
THE DEVELOPMENTAL PLASTICITY OF NEURAL SYSTEMS: IMPLICATIONS FOR PREVENTION SCIENCE
Chair: Elizabeth Ginexi
THE DEVELOPMENTAL PLASTICITY OF NEURAL SYSTEMS: IMPLICATIONS FOR PREVENTION SCIENCE. Elizabeth Ginexi1, Daniel Shaw2, 1National Institutes of Health, Bethesda, MD United States; 2University of Pittsburgh, Pittsburgh, PA United States
Adversity and trauma experienced in early life may compromise social, behavioral, cognitive and emotional functioning in profound ways. Compelling evidence from a variety of animal and human studies suggests that early life stress may increase risk for behavioral and mental disorders including substance use initiation and escalation to abuse. Neurobiological research has indicated that the relationship between early life stress and the development of mental and physical disorders may be mediated by changes in the hypothalamic-pituitary-adrenal (HPA) axis, as well as alterations in other structural and functional neural systems involved in the regulation of stress responses. Individual differences in adaptations to stress may influence the development of cognitive and emotional functions in ways that have relevance to prevention science and practice. This notion is especially pertinent given research on brain plasticity, which suggests that experiences can change neural activity patterns, for the better or for worse. This symposium will highlight how prevention science can translate research findings on the effects of adversity on developing neural systems and on neural plasticity into testable intervention models that may give us a better understanding of how and when to intervene to prevent or reverse these effects. Three panelists will discuss current findings related to the flexibility of specific neural systems known to be affected by early experiences. Dr. Ronald Dahl will present current neuroscience findings related to emotional influences on decision-making in adolescents. Dr. Phil Fisher will discuss research on maltreated children and associated stress reactivity/regulation. Dr. Diana Fishbein will present new findings on the relationship between stress exposure, executive cognitive and emotional deficits, and drug use in the context of a preventive intervention among high-risk adolescents. Discussion and synthesis, led by Dr. Daniel Shaw, will focus on how these related basic neurobiological findings might inform psychosocial models and the development or refinement of preventive interventions to improve the health and well being of youth.
ADOLESCENT BRAIN DEVELOPMENT: A PERIOD OF VULNERABILITIES AND OPPORTUNITIES. Ronald Dahl1, 1University of Pittsburgh, Pittsburgh, PA United States
Adolescent development is a period of special opportunities as well as vulnerabilities with respect to a wide range of behavioral and emotional problems in youth. This presentation considers the role of brain/behavior/social-context interactions during pubertal maturation that can influence these developmental trajectories in positive and negative ways. A model is described that focuses on neurobehavioral changes at puberty that lead to an increased tendency toward risk-taking and sensation-seeking in adolescence. These biologically-based changes in drives, emotions and motivations (“igniting passions”) often emerge early in adolescence, whereas the gradual and relatively prolonged maturation of self-regulatory skills and judgment continue to develop through late adolescence. This may be particularly relevant to adolescents living in social contexts that create challenges to the gradually emerging skills in self-control. Data in support of key aspects of this model will be presented. In particular, studies of emotional influences on decision making will be highlighted. The clinical and social policy implications of this model—and its relevance to prevention science—are also discussed.
PLASTICITY AND CONCORDANCE OF HPA AXIS ACTIVITY AND BEHAVIORAL ADJUSTMENT AMONG FOSTER PRESCHOOLERS IN THE EARLY INTERVENTION FOSTER CARE PROGRAM. Philip Fisher1, Katherine Pears1, 1Oregon Social Learning Center, Eugene, OR United States
Although an extensive literature has documented associations among early life stress, alterations in the functioning of the hypothalamic-pituitary-adrenal (HPA) axis, and problems with behavioral adjustment, this work has been largely post-hoc and descriptive. Less is known about the sensitivity of stress regulatory neural systems to improved environmental conditions and the extent to which behavioral changes that may occur with improved circumstances may be matched with more typical functioning in systems such as the HPA axis. The Early Intervention Foster Care (EIFC) randomized trial has gathered prospective longitudinal data among children starting placement in therapeutic foster care, and thus provides an opportunity to monitor changes in, and the associations between, behavioral adjustment and HPA activity over time. The EIFC program, based on principals of social learning and attachment theories, is delivered via a team approach including intensive consultation and support to foster caregivers, services to children, and parent training for birth parents and long-term placement resources (Fisher, Ellis, & Chamberlain, 1999). Emerging evidence indicates that permanent placements following foster care are significantly less likely to fail for EIFC participants compared to children in regular foster care, regardless of the number of prior placements (Fisher, Burraston, & Pears, in press).
To assess changes in behavioral adjustment and HPA axis activity, we gathered data at monthly intervals from children in the EIFC intervention and foster care comparison (FCC) groups. Measures included caregiver reported behavior problems, attachment behavior, and morning and evening salivary cortisol as an indicators of HPA axis activity. Improvements were observed on a number of behavioral adjustment measures among EIFC condition children. Also, among EIFC children who showed low cortisol levels at baseline, post-intervention levels approached more typical HPA function levels. In contrast, there were few improvements in behavioral functioning among FCC condition children, and the FCC group as a whole showed lower cortisol levels over time. Most notably, FCC children with low cortisol at baseline showed no changes, whereas those with typical cortisol showed significantly decreasing levels. These results indicate that there is both plasticity in HPA function, and concordance between changes in behavioral adjustment and HPA axis function. It is perhaps most significant that these changes appear to have the potential to be both positive (as in the case of EIFC condition children) and negative (as in the case of the FCC children). We will discuss the implications of these findings for policy and programming for high-risk children, and for future research.
UNDERLYING MECHANISMS IN DIFFERENTIAL RESPONSIVITY TO DRUG ABUSE PREVENTIVE INTERVENTIONS AMONG ADOLESCENTS. Diana Fishbein1, Christopher Hyde1, Diana Eldreth2, Mallie J. Paschall3, Robert Hubal4, Abhik Das1, Ralph Tarter5, Nicholas S. Ialongo6, Scott Hubbard6, Betty Yung7, 1Research Triangle Institute, Rockville, MD United States; 2RTI International, Baltimore, MD United States; 3Research Triangle Institute, Chapel Hill, NC United States; 4Research Triangle Institute, Research Triangle Park, NC United States; 5University of Pittsburgh, Pittsburgh, PA United States; 6Johns Hopkins University, Baltimore, MD United States; 7Wright State University, Dayton, OH United States
Developing a portfolio of effective drug abuse preventive interventions requires a determination of how to match particular subgroups with specific programs. It is, thus, critical that we differentiate those who respond favorably to interventions considered to be effective from those who do not on the basis of functional indicators. The executive cognitive functions (ECF) and their modulation of emotional responses may comprise key facets of regulatory processes related to risk for substance abuse. Hence, they may be integral to predicting differential responses to responses to prevention programming. The present study, conducted in collaboration with the ongoing preventive intervention study at Johns Hopkins University Prevention Intervention Research Center (JHU PIRC) within the Baltimore City Public Schools, was directed at elucidating the neurobiological mechanisms associated with substance abuse risk. A subsample (N=120) from this larger study was selected in grade 9, one group with a previous or current diagnosis of Conduct Disorder (CD) and related behavioral problems and the second without CD or any other reported behavioral problems. Extensive longitudinal data (10 years) collected from the child, parents, teachers and peers were available, including information on stress exposures. The participants received a general intelligence test, three developmentally appropriate ECF tasks, and a measure of emotion perception. In a separate session conducted 6-8 weeks later, adolescents were randomly assigned to either an experimental or control group. The experimental group was facilitated using videotaped stimulus materials from a model preventive intervention, namely, the Positive Adolescent Choices Training (PACT). Two behavioral outcome measures recorded change from baseline: 1) a series of virtual reality vignettes reflective of real-life situations involving behavioral choices and, 2) two social cognition questionnaires. Results indicated that poor performance on ECF and emotional perception tasks and a diagnosis of CD each predicted absence of behavioral change in response to the prevention curriculum. Furthermore, the relationship between stress exposures and drug use was mediated by behavioral manifestations of ECF abilities. This study provides valuable insights for designing individualized prevention interventions that take into account psychopathology and its neurological substrate for adolescents who may be refractory to standard substance abuse preventions.
CONCURRENT 3, EPIDEMIOLOGY, Poster Forum
MULTIPLE PROBLEMS OF YOUTH AND YOUNG ADULTS, PEER GROUPS AND RISK BEHAVIORS, AND PHYSICAL AGGRESSION AMONG UNIVERSITY STUDENTS
Chair: Robert Saltz
ILLICIT DRUG USE, PROBLEM DRINKING, AND RELATED PROBLEMS IN ADOLESCENTS AND YOUNG ADULTS. Meng-Jinn Chen1, 1Prevention Research Center, Berkeley, CA United States
Prior research indicates that young people's use of marijuana and other illicit drugs is closely associated with the use of cigarettes and/or problem drinking and few illicit drug users use only illicit drugs (Mohler-Kuo, Lee, & Wechsler, 2003; Sutherland & Willner, 1998). Use of alcohol and illicit drugs in combination is associated with more negative consequences than use of alcohol alone (Bennett, Miller, & Woodall, 1999; Sutherland & Willner, 1998). The present study investigated the problem of poly-substance use among adolescents and young adults. The data were collected using self-administered paper and pencil questionnaire surveys and from an ethnically diverse sample of 1,044 community college students in the central valley of California (aged 15-25 [mean=18.9, SD=1.81]; 43% male; 39% White, 27% Latino, 20% Asian, 5% Black, and 9% other). Students were asked their use of substances in the past 12 months and use of cigarettes and illicit drugs while they were drinking alcohol. Descriptive statistics showed that 39% of the subjects used marijuana in the past 12 months, 14% used club drugs, 8% used narcotics or cocaine, 36% smoked cigarettes, and 73% drank alcohol. Consistent with prior research, few illicit drug users used only illicit drugs. Ninety-six percent of students who reported illicit drug use during the past 12 months also drank alcohol during those 12 months and 82% reported using illicit drugs while they were drinking alcohol. Use of alcohol and illicit drugs in combination was more closely associated with drinking occasions involving either small or large groups of people (r = .32 and .30, respectively, ps < .001) than with drinking alone (r = .15, p < .001). Logistic regression analyses controlling for age and gender showed that students who used illicit drugs while they were drinking were much more likely to engage in problem behaviors than those who used only alcohol. The risks for intoxication, driving under the influence of alcohol, and riding with a driver who had been drinking were 4.0 times, 3.3 times, and 2.3 times greater, respectively. Further, students who used illicit drugs while they were drinking alcohol were 5.6 times as likely to meet criteria for problem drinking based on the AUDIT threshold score (8 and greater) as those who used only alcohol. Implications for prevention efforts that target young people's substance use and related problems are discussed.