PROGRAM IMPLEMENTATION FIDELITY: FINDINGS FROM A SCHOOL-BASED PEER-LED CURRICULUM. Valerie Johnson1, Sherry Barr2, Robert Pandina1, Sharon Rose Powell2, John Kalafat1, 1Rutgers, The State University of New Jersey, Piscataway, NJ United States; 2Princeton Center for Leadership Training, Princeton, NJ United States
While drug abuse prevention researchers have generally concluded that the success of a program depends, in part, on the fidelity of implementation, the literature on studies of implementation fidelity has been limited (Dane & Schneider, 1998; Dusenbury et al., 2003).
In the fall of 2004, researchers and practitioners from the NIDA funded Rutgers Transdisciplinary Research Center initiated the pilot year of a peer led prevention curriculum in an urban high school with more than 90% of students being of minority status. This curriculum consists of trained teacher-advisors who provide upperclass students with a daily class to instruct them in conducting weekly outreach sessions with freshmen students.
The sessions are comprised of activities that help students in the transitional period of moving into high school and focus upon important risk or protective factors found to be salient in the development of substance use and related problems. These factors include peer and school bonding, academic motivation, normative expectations surrounding use, availability and utilization of resources, as well as time, stress and anger management, to name a few.
As part of the research design, an extensive protocol of self report and observational data were collected over the school year which assessed 1) level of adherence to the program, 2) quality of program delivery, and 3) participant responsiveness. Multiple forms with a variety of measures were collected at two teacher led class periods per week and from the weekly freshmen outreach sessions.
In addition to reporting on the above outlined information, we will describe the reliability of our measures and compare and contrast the self-report and observation methodologies. We will discuss findings that may aid in the specification of criteria for effective implementation; including teacher and peer leader qualifications and training outcomes, leader-student ratios, frequency and duration of the sessions and the particular sessions that were most valued.
This research is supported by NIDA 017552 & Peter F. McManus Charitable Trust.
EXAMINING THE IMPACT OF PARENT AND YOUTH PARTICIPATION RATES ON OUTCOME EFFECTS OF THE COPING POWER PROGRAM. Caroline Boxmeyer1, John Lochman1, Nicole Powell1, Michael Windle2, 1The University of Alabama, Tuscaloosa, AL United States; 2University of Alabama at Birmingham, Birmingham, AL United States
Background: Intent-to-treat analysis has traditionally been the gold standard analytic approach for examining intervention effects in randomized controlled trials. However, this approach may limit what could be learned about the true effect sizes of interventions. In particular, inclusion of individuals who receive a low “dose” of the intervention in outcome analyses may lead to nonsignificant intervention effects, or to a reduction in the overall effect size in the presence of statistically significant outcomes. This methodological issue will be illustrated utilizing data from a recent evaluation of the Coping Power Program, a school-based preventive intervention for aggressive children and their parents.
Methods: 160 fifth grade boys and girls identified as at-risk for delinquent behavior were assigned to either the Coping Power intervention (n=80) or a control condition (n=80). The child component of the Coping Power intervention included 24 group sessions. The parent component consisted of 10 group sessions. Intervention effects were assessed by obtaining parent and teacher ratings of externalizing behavior problems on the Behavior Assessment System for Children (Reynolds & Kamphaus, 1992).
Results: Parents who received a relatively full dose of the Coping Power parent group sessions (i.e., attended at least half of the sessions; n=27), rated their children as having significantly greater reduction in externalizing behavior problems than did parents who attended fewer than half of the sessions (n=40; F=4.14, p<.05). Since nearly two-thirds of parents in the intervention condition received a relatively low dose, this may have contributed to the lack of a significant overall intervention effect for parent-reported externalizing behavior problems. There was a significant overall intervention effect for teacher-reported externalizing behavior problems. However, closer inspection of the data indicated that children who did not receive the full Coping Power group intervention did not exhibit significant reductions in externalizing behavior problems relative to children in the control condition, while children who received the full group intervention did.
Discussion: These findings illustrate how inclusion of individuals who receive low dosage of an intervention may mask or underestimate true intervention effects, which can have a profound impact on policy decisions for prevention and treatment of children´s mental health problems. This presentation includes a discussion of alternative strategies for incorporating the amount of exposure to an intervention in outcome analyses while controlling for potential confounds (e.g., LATE, CACE; Dunn et al., 2003) as well as strategies for enhancing parent and youth participation in preventive interventions.
MOTIVATIONAL CHANGES FOR PARTICIPATION IN EXPERIMENTAL HIV PREVENTION INTERVENTIONS. Carl Latkin1, Amy Knowlton1, 1Johns Hopkins University, Baltimore, MD United States
Background: Few studies have examined participants´ motivations for enrolling in behavioral prevention interventions. The present study utilized quantitative and qualitative methods to examine initial motives and changes in motives during the course of an experimental small group behavioral HIV prevention intervention.
Methods: The intervention was a RCT that targeted HIV risk behaviors among African American injection drug users in an inner-city setting and trained them to promote risk reduction among their network members. 838 follow-up interviews were conducted, and themes were derived from 30 qualitative interviews and observations and tape recordings of 18 intervention sessions.
Results: Among this sample, initial primary motives for participation included financial remuneration, desire for social services, and control of drug use. However, through the course of the intervention, intrinsic motivations became more salient. These include desire to improve communication skills, gain greater respect in the community through peer outreach education, and affirmations by other intervention group members. Quantitative analyses revealed that drug users were often more concerned about being mugged, paying bills, and obtaining employment than about becoming infected with HIV.
Conclusion: Results from this study suggest that motives to participate in an HIV prevention interventions change over time and that for many participants HIV prevention is less of a concern than other salient life stressors.
THE EFFECTS ON HIGH SCHOOL STUDENTS OF USING THEM TO DELIVER AN ALCOHOL PREVENTION PROGRAM TO ELEMENTARY STUDENTS. Mary Lou Bell1, Alison Padget1, 1Bell Group, Austin, TX United States
We present the results of a controlled evaluation of the effects on high school students of using them to teach Protecting You/Protecting Me (PY/PM), a classroom-based alcohol prevention and SAMHSA model program developed my Mothers Against Drunk Driving (MADD), to elementary school students. We surveyed high school students enrolled in a peer helping course (N=218) before we trained them in PY/PM and again after they taught the curriculum to elementary school students during the 2003-2004 school year. We surveyed similarly selected peer helpers (N=183) from matched comparison schools at the same times. The survey included measures of 30-day alcohol use, knowledge of the effects of alcohol, perceptions of the risks of low and high alcohol use, self-efficacy, attitudes toward future drinking, and attitudes toward teaching children. We obtained matched pre- and post-tests for 191 and 142 peer helpers in the intervention and comparison groups, respectively. To determine differences on key outcomes between our intervention and comparison groups we conducted multiple regression analyses that controlled for students´ race, gender, and pre-test score differences. Relative to the comparison group, students in the intervention group made gains in the areas of understanding the effects of underage alcohol use, attitudes toward future drinking, knowledge of the risks of high levels of alcohol consumption, and attitudes toward teaching children. Intervention students were also less likely at post-test to have consumed wine, consumed more than 5 alcoholic beverages in one sitting, or driven within two hours of drinking in the past 30 days. There were no significant differences in self-efficacy or perceived risks of low levels of alcohol use. Evaluation results demonstrate the potential effectiveness of using high school students as educators in prevention efforts for younger children.
EFFECTS OF VIOLATING THE HOMOGENEITY ASSUMPTION ON DETECTING THE PROGRAM EFFECT OF THEORY-BASED MULTILEVEL PREVENTION PROGRAMS. Oi-Man Kwok1, Jenn-Yun Tien1, Myeongsun Yoon1, 1Arizona State University, Tempe, AZ United States
There is a steady increment of applying the multilevel modeling (MLM) to analyze the data with multilevel structure from theory based prevention studies. However, most of these studies have assumed homogeneity (i.e., homogeneous residual variances) between different subgroups (e.g., treatment/control conditions, and gender) at different levels (e.g., between-family level vs. within-family level). The effect of violation of the homogeneity assumption under the ordinary least square (OLS) regression has been intensively studied in the last decade (Aguinis, 2004). Ignoring the existence of heterogeneity under the OLS regression results in bias estimations of the standard errors of the interested effects (e.g., program effect), which in turn, leads to either inflation of the nominal alpha level or reduction in statistical power for detecting the interested effects. Up to date, very few studies applying the MLM have examined the homogeneity assumption. In this study, we will focus on a two-level model (e.g., participants nested within intervention groups) with prevention program implemented at the higher level (i.e., group level). Monte Carlo study will be conducted to examine the effect of ignoring heterogeneity at different levels on estimation of the program effect and the corresponding standard error, with considering different sample sizes and the extents of the residual variances at different levels. An empirical data will be adopted for illustration.
LEARNING FROM SUCCESSFUL PREVENTION EFFORTS: HIV/AIDS PREVENTION IN SENEGAL. Khadidiatou Ndiaye1, 1Pennsylvania State University, University Park, PA United States
In a world of countless technological progress and amid numerous innovations, AIDS is still devastating lives with no signs of slowing down. In the absence of a cure, communication leading to prevention and understanding of the disease has been the only available weapon. In the midst of widespread despair, the successes of African countries such as Uganda and Senegal in combating HIV/AIDS are rare positive stories.
The present study seeks to understand the communication strategies guiding the success of HIV/AIDS prevention in Senegal. Grounded theory is used as the framework for analysis in exploring the two following research questions: (1) What are the communication strategies of HIV/AIDS prevention programs in Senegal? (2) What are the organizational dimensions of HIV/AIDS prevention programs in Senegal?
The findings revealed four communication strategies used in HIV/AIDS prevention programs in Senegal: (1) Building and reinforcing existing communication structures, (2) using a maximized communication model, (3) designing communication messages that are socially and culturally appropriate, and (4) using communication guided by field research.
At the organizational level, the study discovered that the HIV/AIDS prevention organizations in Senegal were characterized by: (1) Clearly defined structures and diverse memberships, (2) connectedness in fighting the HIV/AIDS epidemic, and (3) openness.
In the discussion chapter, the findings were overviewed to reveal key elements explaining the low HIV prevalence rates in Senegal. Multiplicity, empowerment, leadership, and consciousness were the key factors of effective HIV/AIDS prevention identified from the discussion of the interviews and the documents reviewed. Finally, a framework for effective HIV/AIDS was developed based on the lessons learned in the Senegalese context.
CONCURRENT 4, CULTURAL SENSITIVITY, Roundtable
MOVING TOWARDS STATE OF THE ART, CULTURALLY RELEVANT PREVENTION INTERVENTIONS FOR MINORITY YOUTH
Chair: Lori Holleran
MOVING TOWARDS STATE OF THE ART, CULTURALLY RELEVANT PREVENTION INTERVENTIONS FOR MINORITY YOUTH. Lori Holleran1, Stephanie Coard2, Karol Kumpfer3, Michael Hecht4, Howard Stevenson5, Vanessa Nyborg6, 1University of Texas at Austin, Austin, TX United States; 2Duke University, Durham, NC United States; 3University of Utah, Salt Lake City, UT United States; 4Pennsylvania State University, University Park, PA United States; 5University of Pennsylvania, Philadelphia, PA United States; 6University of California, Santa Barbara, Santa Barbara, CA United States
Ethnic minorities experience a disproportionate amount of unmet mental health needs, negatively impacting their health and functioning. The Surgeon General´s 2001 report outlines the mental health burdens, barriers to care, and recommendations for improving care among ethnic minorities. One recommendation calls for increased research on the role of race, ethnicity and culture in mental health, and additional research on the utility of culturally-adapted approaches for preventing/reducing stigma and increasing utilization of interventions/services. Some researchers suggest that culturally-adapted interventions increase involvement and retention of ethnic minorities, but have not improved outcomes (Kumpfer et al., 2002). Others believe cultural adaptations can improve outcomes. The prevention field remains divided and the extent to which cultural modifications improve the efficacy of interventions beyond standard evidenced-based program outcomes requires examination. Few randomized controlled trials have adequately addressed this issue and many questions remain: How are ethnic/cultural constructs understood/considered in the development of preventive interventions? What are the challenges/benefits of doing so? Who is best served? What are public health implications?This roundtable discussion/scientific dialogue addresses the role of culture in the development and implementation of prevention interventions. A diverse panel (comprised of junior and senior level prevention researchers) will provide perspectives on interventions for African American, Latino and Native American groups. Each presenter will briefly review their intervention approach, with attention to the role of culture within their models, and issues of theory, translation, measurement and implementation. In doing so, panelists will be asked to respond to the above questions among others. The co-chairs, Lori Holleran, Ph.D. and Stephanie Coard, Ph.D., both early career investigators (K01 Awardees from NIH), will summarize the different viewpoints and facilitate dialogue among session panelists and participants to further understanding of the role of culture in prevention research, and identify and address the challenges in developing culturally appropriate interventions.Discussant panelists are: Karol Kumpfer, Ph.D.: cultural adaptations of evidence-based family programs for Native Americans and Asian/Pacific Islanders; Michael Hecht, Ph.D.: culturally grounded substance use prevention with Latino and African American youth; Howard Stevenson, Ph.D.: protective role of racial identity and racial socialization processes in prevention efforts with African-American students and families; and Vanessa M. Nyborg, Ph.D.: Culturally appropriate preventive interventions for African American boys.
CONCURRENT 5, ETIOLOGY, Organized symposia
MALTREATMENT, RISK AND RESILIENCE: FINDINGS FOR PUBLIC POLICY
Chair: Miriam Ehrensaft
MALTREATMENT, RISK AND RESILIENCE: FINDINGS FOR PUBLIC POLICY. Miriam Ehrensaft1, 1Columbia University, New York, NY United States
Research has long supported an increased risk for emotional and behavior problems among children who experience maltreatment, including abuse, neglect, and frequent corporal punishment. Further longitudinal research is needed to identify pathways to adverse outcomes among children who experience maltreatment, as well as factors that may promote resilience following the experience of maltreatment. Such basic studies are critical to the design of empirically informed public policy and interventions for maltreated children. The present symposium integrates three studies of risk and protective factors for maltreatment. The first study, based on a community sample of children followed for over 25 years, examines the effects of parental corporal punishment on child behavior problems across two generations, testing whether and why the effects of using corporal punishment on subsequent child behavior problems have changed across two generations. The second study aims to identify, in a large, nationally representative sample of young twin pairs, individual, family, and neighborhood factors that distinguish children who were resilient to maltreatment from those who were not. This study also pays special attention to role of individual child factors, such as temperament and IQ, on resilience to maltreatment. The third study tests a prospective model of attention problems in sexually abused children, highlighting the roles of dissociation and the child´s relationship to the abuse perpetrator. Each of these three studies aims to generate empirical findings that may inform the translation of basic science to public health policy and practice. The discussant for this symposium will integrate the findings of these three studies, and identify implications for public health policy.
INDIVIDUAL, FAMILY, & NEIGHBORHOOD FACTORS THAT PROMOTE RESILIENCE TO MALTREATMENT. Sara Jaffee1, 1University of Pennsylvania, Philadelphia, PA United States
Children who are maltreated are at risk for a range of problems in childhood, adolescence and adulthood However, not all children who are maltreated experience these difficulties. These resilient children achieve positive developmental outcomes despite the significant adversities they have experienced. The goal of the current study was to identify individual, family, and neighborhood factors that distinguished children who were resilient to maltreatment from those who were not and to test hypotheses about the conditions under which children's strengths would predict resilience to maltreatment.
Data were from the Environmental Risk Longitudinal Study which followed a nationally-representative sample of 1,116 twin pairs and their families in the United Kingdom from age 5 to 7 years. When children were 5 years old, mothers were asked to report whether children had ever been physically maltreated. Parents and children also provided information about characteristics of the children (temperament, IQ), characteristics of parents and families (parent psychopathology, social deprivation), and characteristics of their neighborhoods (crime, social cohesion and control). Children who were characterized as “resilient” met the following criteria: (a) they had been maltreated by age 5 years, and (b) when they were 5 and 7 years of age their teachers reported that their externalizing problems were no greater, on average, than those of children who had never been maltreated. Resilient children could not be distinguished from non-maltreated children with respect to internalizing symptoms, reading ability, or prosocial behavior.
The results of multinomial logistic regression analyses showed that child, family, and neighborhood characteristics differentiated resilient from non-resilient children. Boys who were of above-average intelligence and who were temperamentally sociable and behaviorally-controlled were more likely to be resilient versus non-resilient. Children whose parents had clinically-significant depressive symptomatology and alcohol problems were less likely to be resilient versus non-resilient. Finally, children who lived in neighborhoods that were high in crime and low in cohesion and informal social control were less likely to be resilient versus non-resilient. We tested the hypothesis that individual characteristics would fail to predict resilience when children were exposed to multiple social stressors in the family and neighborhood. Support for this hypothesis was not found. Characteristics like high IQ and well-adjusted temperament predicted resilience to maltreatment even when children were experiencing multiple stressors aside from maltreatment.
A PROSPECTIVE MODEL OF ATTENTION PROBLEMS IN SEXUALLY ABUSED CHILDREN. Julie Kaplow1, Erin Hall2, Karestan Koenen3, Lisa Amaya-Jackson4, 1University of Medicine and Dentistry of New Jersey, Newark, NJ United States; 2Boston University, Boston, MA United States; 3Harvard School of Public Health, Boston, MA United States; 4Duke University Medical Center, Durham, NC United States
Whereas one of the most frequently diagnosed disorders in victims of child sexual abuse (CSA) is ADHD, systematic research regarding predictors of later attention problems in this population is lacking. The goals of the current study are to develop and test a prospective model of attention problems in sexually abused children including pretrauma, trauma, and disclosure-related variables. Participants in the current study had been referred to a treatment facility that offers services to child victims of sexual abuse. The sample was comprised of 129 girls and 27 boys, with a mean age of 10.7 years. Of the sample, 56% were African American, 23% were Caucasian, 12% were Native American, 5% were Bi-racial, and 4% were Hispanic.Information regarding gender, age of sexual abuse onset, and relationship to perpetrator were gathered from written reports of forensic interviews at Time 1. In addition, dissociation and PTSD symptoms were assessed at Time 1 by the Dissociation and PTSD scales of the Trauma Symptom Checklist for Children (TSCC), a 54-item self-report instrument for children. Attention problems were assessed at Time 2 through the Attention Problems scale of the Child Behavior Checklist for Children (CBCL), a parent-report measure of the child's internalizing and externalizing symptomatology. The Things That You Have Done Survey is a self-report measure used to assess delinquent behavior in children. Areas comprising the General Delinquency Scale of this measure include physical aggression, stealing, vandalism, and substance abuse, which were also assessed at Time 2. The final path analytic model provided excellent fit indices, using the Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Chi Square (chi 2): CFI = 1.00, RMSEA = .00, chi 2 = 10.16, df = 12, p = .60. The results indicate two direct paths from trauma and disclosure-related variables [relationship to perpetrator (beta = .30) and dissociation (beta = .41)] to later attention problems, while controlling for concurrent delinquency (beta = .32). Interestingly, PTSD was only indirectly associated with attention problems via dissociation. Taken together, these pathways accounted for approximately 37% of the variance in attention problems. The findings suggest that children who were sexually abused by someone in their family and/or exhibited dissociative symptoms at the time of their disclosure are at increased risk for developing later attention problems. These findings have important implications for the assessment and referral of sexually abused children as well as the prevention of attention problems in this population.