|I.||Current Issues and Themes in Prevention Research||
|II.||The Process of Program Review||
|III. Universal Preventive Interventions||
|IV. Programs That Focus on Externalizing Behaviors||
|V. Programs That Focus on Internalizing Behaviors||
|VI. Summary of Findings||
|Adolescent Transitions Program||66|
|Anger Coping Program||69|
|Attributional Intervention (Brainpower Program)||74|
|Big Brothers/Big Sisters||76|
|Child Development Project||78|
|Children of Divorce Intervention Program||80|
|Children of Divorce Parenting Program||83|
|Coping With Stress Course||86|
|Counselors CARE and Coping and Support Training||88|
|Earlscourt Social Skills Group Program||91|
|Family Bereavement Program||93|
|First Step to Success||97|
|Good Behavior Game||99|
|Improving Social Awareness-Social Problem Solving||102|
|Interpersonal Cognitive Problem Solving||104|
|Intervention Campaign Against Bully-Victim Problems||106|
|Linking the Interests of Families and Teachers||108|
|Montreal Longitudinal Experimental Study||111|
|Peer Coping Skills Training||114|
|Penn Prevention Program||116|
|Positive Youth Development Program||119|
|Promoting Alternative THinking Strategies||121|
|Primary Mental Health Project||124|
|Queensland Early Intervention and Prevention Project||127|
|Responding in Peaceful and Positive Ways||129|
|School Transitional Environment Project||131|
|Seattle Social Development Project||133|
|Second Step: A Violence Prevention Curriculum||136|
|Social Relations Program||138|
|Stress Inoculation Training I||140|
|Stress Inoculation Training II||143|
|Suicide Prevention Program I||145|
|Suicide Prevention Program II||147|
This report is the result of a contract awarded by The Center for Mental Health Services to the Prevention Research Center for the Promotion of Human Development at Pennsylvania State University.
Goals of the Review and Report
The central goal of this report is to review and summarize the current state of knowledge on the effectiveness of preventive interventions intended to reduce the risk or effects of psychopathology in school-age children. In doing so, this report
The current report is not intended to describe, in detail, preventive interventions that show effects only on outcomes such as substance use, sexual behavior and contraception, or interventions that promote competence, but have not demonstrated effects on psychological symptomology. A number of recent reports review these related, but separate fields (Catalano, et al, 1998; Durlak, 1995; Durlak and Wells, 1998; Greenberg, Zins, Elias, & Weissberg, 1999; Kirby, et al., 1994; Tobler and Stratton, 1997).
Structure of the Report
The report contains six sections as well as Appendices. Section One reviews current issues and themes in prevention research with school-aged children. Section Two summarizes the process of program review. Section Three reviews universal prevention programs. Sections Four and Five review selective and indicated programs for children at risk for either externalizing or internalizing psychopathology. Section Six provides a brief summary of the findings of this review and their implications for program development and evaluation. Finally, the Appendices provide in-depth information on the effective programs identified in Sections Two, Three and Four.
In the last decade prevention has moved into the forefront and become a priority for many federal agencies in terms of policy, practice, and research. This paradigm shift began with a report by the National Advisory Mental Health Council (1990) and is reflected in the combined work of the National Institute of Mental Health (NIMH, 1993) and the Institute of Medicine (IOM, 1994). More recently, the National Advisory Mental Health Council Workgroup on Mental Disorders Prevention Research (NIMH, 1998) outlined a number of priorities and recommendations for research initiatives in prevention science.
The Need for a Preventive Focus in Child Mental Health
Interest in prevention is also reflected in the goals that have been set for our nation's health. One of the original objectives of Healthy People 2000 was to reduce the prevalence of mental health disorders in children and adolescents to less than 17%, from an estimated 20% among youth younger than 18 in 1992 (DHHS, 1991). As of 1997, the summary list of mental health objectives for Healthy People 2000 included reducing suicides to no more than 8.2 per 100,000 youth (aged 15-19) and reducing the incidence of injurious suicide attempts among adolescents to 1.8% and, more specifically, to 2.0% among female adolescents (DHHS, 1997). A number of other objectives were related to child and adolescent mental health. One of the risk reduction objectives in the Violent and Abusive Behavior category was to reduce the incidence of physical fighting among adolescents aged 14-17 from a baseline of 137 incidents per 100,000 high school students per month to 110 per 100,000 (DHHS, 1997). Two additional objectives in this category were to increase to at least 50% the proportion of elementary and secondary schools that include nonviolent conflict resolution skills and to extend violence prevention programs to at least 80% of local jurisdictions with populations over 100,000 (DHHS, 1997). It is unlikely that these goals will be met by the year 2000.
There is growing concern in our country as increasing numbers of children and adolescents are having difficulty managing the challenges of development. Between 12% and 22% of Americas youth under age 18 are in need of mental health services (National Advisory Mental Health Council, 1990), and an estimated 7.5 million children and adolescents suffer from one or more mental disorders (OTA, 1986). In addition to the personal suffering experienced by children with emotional or behavioral problems and their families, mental health disorders also have a tremendous cost to society. According to the National Advisory Mental Health Council (1990), in 1990 mental illness cost the United States an estimated 74.9 billion dollars.
While a number of recent reviews (e.g. Kazdin & Weisz, 1998; or see special issue of Journal of Clinical Child Psychology, 27, 1998) and meta-analyses (Casey & Berman, 1985; Kazdin, Bass, Ayers, & Rogers, 1990; Weisz, Weiss, Alicke, & Klotz, 1987; Weisz, Weiss, Han, Granger, & Morton, 1995) provide evidence that childhood disorders are amenable to treatment, the literature must be interpreted cautiously. There is still a great deal to be learned about specific types of treatments, their appropriateness for certain disorders, and the factors that contribute to treatment success and failure. We have not reached the point where we are able to serve all children effectively. As suggested by the Institute of Medicine in their report to Congress on the state of prevention research in mental health, it is important not to overlook the significance of prevention even if treatment efforts have been unsuccessful; in fact, prevention may play a particularly important role for these types of disorders (IOM, 1994).
It is clear that to reduce levels of childhood mental illness, interventions need to begin earlier, or ideally, preventive interventions need to be provided prior to the development of significant symptomology. In addition, efforts need to be increased to reach the many children that do not have access to treatment. Many children and adolescents with clinical levels of problems never receive appropriate mental health services or they receive inappropriate services (Knitzer, 1985; Tuma, 1989). Another problem with service delivery is that some children only become eligible for therapeutic services after they have entered another system such as special education or juvenile court and this is usually after their problems have begun to escalate.
The Role of Developmental Theory in Prevention Research
Prevention science is highlighted by the integration of developmental theory with models from public health, epidemiology, sociology, and developmental psychopathology in conceptualizing, designing, and implementing preventive interventions (Cicchetti, 1984; Cicchetti & Cohen, 1995; Kellam & Rebok, 1992; Lorion, 1990; Sameroff, 1991; Sroufe & Rutter, 1984). As concepts in development have broadened to include ecological analysis (Bronfenbrenner, 1979, 1995; Garbarino, 1992) and multivariate examination of causation and risk (Institute of Medicine, 1994; Rutter, 1987), developmental theory has provided a powerful framework for organizing and building the field.
Given the principle that the developing organism is strongly influenced by context, Bronfenbrenner's model of the nature and levels of context has catalyzed the field (Bronfenbrenner, 1979; Bronfenbrenner & Crouter, 1983; Bronfenbrenner, 1995). The ecological model posits four levels for classifying context beginning with those ecologies in which the child directly interacts and proceeding to increasingly distant levels of the social world that affect child development. The first level, the microsystem, is composed of ecologies with which the child directly interacts such as the family, school, peer group, and neighborhood. The mesosystem encompasses the relationships between the various microsystems (e.g., the family-school connection or between the parents and the child's peer group and peers' families). The absence of mesosystem links may also be an important risk factor in development.
Interactions within both the microsystem and mesosystem are often affected by circumstances that do not directly involve the child. For example, children and youth may be significantly affected by changes in marital circumstance, parental social support, changes in the legal system (e.g., changing definitions of neglect or abuse; regulation of firearms, tobacco, and illegal drugs), the social welfare system (e.g., welfare reforms, boundary changes for categorical services), the mass media (e.g., controls on children's exposure to television violence, the widened horizons via the internet), or other social structures that set policies and practices that alter microsystem and mesosystem interactions. The exosystem is those contexts and actions that indirectly impact the child's development. Many preventive interventions may be viewed as changes at the exosystem level that alter interactions among lower system levels. Finally, the macrosystem represents the widest level of systems influence, consisting of the broad ideological and institutional patterns and events that define a culture or subculture.
Developmental-ecological models can be used both to frame basic research attempts to understand layers of influence on behavior, and also to identify potential targets and mediators of intervention. It is important for researchers to specify, for example, whether their interventions focus primarily on: the microsystem--or a particular portion of it; multiple microsystems (e.g., interventions for both the home and school); the mesosystem (e.g., the family-school connection); informal networks that in turn affect the microsystem (e.g., the development of extended family or peer support to parents); or developing new models of service delivery or regulatory reform (e.g., formal services in the exosystem). Further, one might ask if these different levels of intervention emphasize changing the behavior and attitudes of individuals at these levels (i.e., person-centered), or changing the nature of the system's operation itself (i.e., environment-focused) (Cowen, 1977; Weissberg, Caplan, & Harwood, 1991).
The Role of Risk and Protective Factors in Preventive Interventions
Public health models have long based their interventions on reducing the risk factors for disease or disorder as well as promoting processes that buffer or protect against risk. Community-wide programs have focused on reducing both environmental and individual behavioral risks for both heart and lung disease and have demonstrated positive effects on health behaviors as well as reductions in smoking (Farquhar et al., 1990; Jacobs et al., 1986; Pushka et al., 1989).
Risk factors and their operation During the past decades, a number of risk factors have been identified that place children at increased risk for psychopathology. Coie et al. (1993, p. 1022) grouped empirically derived, generic risk factors into the following seven individual and environmental domains:
1. Constitutional handicaps: perinatal complications, neurochemical imbalance, organic handicaps, and sensory disabilities;
2. Skill development delays: low intelligence, social incompetence, attentional deficits, reading disabilities, and poor work skills and habits;
3. Emotional difficulties: apathy or emotional blunting, emotional immaturity, low self-esteem, and emotional disregulation;
4. Family circumstances: low social class, mental illness in the family, large family size, child abuse, stressful life events, family disorganization, communication deviance, family conflict, and poor bonding to parents;
5. Interpersonal problems: peer rejection, alienation, and isolation;
6. School problems: scholastic demoralization and school failure;
7. Ecological risks: neighborhood disorganization, extreme poverty, racial injustice, and unemployment.
Theory and research support a number of observations about the operation of these risk factors and the development of behavioral maladaptation. First, development is complex and it is unlikely that there is a single cause of, or risk factor for, any disorder. It is doubtful that most childhood social and behavioral disorders can be eliminated by only treating causes that are purported to reside in the child alone (Rutter, 1982). Furthermore, there are multiple pathways to most psychological disorders. That is, different combinations of risk factors may lead to the same disorder and no single cause may be sufficient to produce a specific negative outcome (Greenberg, Speltz, & DeKlyen, 1993). In addition, risk factors occur not only at individual or family levels, but at all levels within the ecological model (Kellam, 1990).
The complexity of developmental pathways is clear from research relating risk factors to disorders. There appears to be a non-linear relationship between risk factors and outcomes. Although one or two risk factors may show little prediction to poor outcomes, there are rapidly increasing rates of disorders with additional risk factors (Rutter 1979; Sameroff, Seifer, Barocas, Zax, & Greenspan, 1987). However, not all children who experience such contexts develop adjustment problems (e.g. Cowen et al., 1992), and no one factor alone accounts for children's adjustment problems (e.g., Sameroff & Seifer, 1990).
Given the above findings, it is apparent that many developmental risk factors are not disorder-specific, but may relate instead to a variety of maladaptive outcomes. The notion of generic and inter-related risk factors has led to a strategy of targeting multiple factors simultaneously with the hope that the potential payoff will be greater than a focused attack on controlling a single risk factor. Recent findings in behavioral epidemiology indicate that mental health problems, social problems, and health-risk behaviors often co-occur as an organized pattern of adolescent risk behaviors (Donovan, Jessor & Costa, 1988; Dryfoos, 1990; Elliott, Huizinga, & Menard, 1989; Jessor et al., 1991; Jessor & Jessor, 1977). Thus, because risk factors may predict multiple outcomes and there is great overlap among problem behaviors, prevention efforts that focus on risk reduction of interacting risk factors may have direct effects on diverse outcomes (Coie et al., 1993; Dryfoos, 1990).
Protective factors and their operation Protective factors are variables that reduce the likelihood of maladaptive outcomes under conditions of risk. Although less is known about protective factors and their operation (Rutter, 1985; Kazdin, 1991; Luthar, 1993), at least three broad domains of protective factors have been identified. The first domain includes characteristics of the individual such as cognitive skills, social-cognitive skills, temperamental characteristics, and social skills (Luthar & Zigler, 1992). The quality of the child's interactions with the environment comprise the second domain. These interactions include secure attachments to parents (Morissett, Barnard, Greenberg, Booth, & Speiker, 1990) and attachments to peers or other adults who engage in positive health behaviors and have prosocial values. A third protective domain involves aspects of the mesosystem and exosystem, such as school-home relations, quality schools, and regulatory activities. Similar to risk factors, some protective factors may be more malleable and thus, more effective targets for prevention.
Coie et al. (1993) suggested that protective factors may work in one or more of the following four ways: directly decrease dysfunction; interact with risk factors to buffer their effects; disrupt the mediational chain by which risk leads to disorder; or prevent the initial occurrence of risk factors. By specifying links between protective factors, positive outcomes, and reduced problem behaviors, prevention researchers may more successfully identify relevant targets for intervention (Coie et al., 1993; Dryfoos, 1990).
The specification of intervention goals is an important component of preventive-intervention research and practice. This requires both an understanding of risk and protective factors that contribute to outcomes, and also the identification of competencies that are presumed mediators or goals of the intervention. Although these goals may include the prevention of difficulties (e.g., absence of psychopathology, abstention from substance use), they also involve the promotion of healthy developmental outcomes (Pittman & Cahill, 1992). Further, the prevention of deleterious outcomes involves the enhancement of competency mediators (e.g., effective social problem-solving as a mediator of reductions in delinquency).
Preventive Intervention: Definition of Levels
The IOM Report (1994) clarified the placement of preventive intervention within the broader mental health intervention framework by differentiating it from treatment (i.e., case identification; standard treatment for known disorders) and maintenance (i.e., compliance with long-term treatment to reduce relapse; after-care, including rehabilitation). Based, in part, on Gordon's (1983, 1987) proposal to replace the terms primary, secondary, and tertiary prevention, the IOM Report defined three forms of preventive intervention: universal, selective, and indicated.
Universal preventive interventions target the general
public or a whole population group that has not been identified
on the basis of individual risk. Exemplars include prenatal care,
childhood immunization, and school-based competence enhancement
programs. Because universal programs are positive, proactive,
and provided independent of risk status, their potential for stigmatizing
participants is minimized and they may be more readily accepted
and adopted. Selective interventions target individuals
or a subgroups (based on biological or social risk factors) whose
risk of developing mental disorders is significantly higher than
average. Examples of selective intervention programs include:
home visitation and infant day care for low-birth weight children,
preschool programs for all children from poor neighborhoods, and
support groups for children who have suffered losses/traumas.
Indicated preventive interventions target individuals who
are identified as having prodromal signs or symptoms or biological
markers related to mental disorders, but who do not yet meet diagnostic
criteria. Providing social skills or parent-child interaction
training for children who have early behavioral problems are examples
of indicated interventions.
Criteria for Review
Outcomes of Interest. The scope of interest included universal, selective or indicated prevention programs that were found to produce improvements in specific psychological symptomology or in factors generally considered to be directly associated with increased risk for child mental disorders. As such, studies were included in which children showed early problems or high-risk for later disorder, but studies were excluded in which children were given diagnostic interviews and met criteria for DSM-III R or DSM-IV disorders. The age focus was restricted to children from ages 5 to 18.
Programs were excluded if they produced outcomes solely related to substance abuse, sexuality or health promotion, but did not show reductions in symptomology related to mental disorders. However, if prevention or health-promotion programs showed multiple effects that included reduction in psychiatric symptoms, they were included in the review. Given the common comorbidity and shared risk factors of mental health problems with other poor outcomes such as delinquency and substance abuse, the lines of distinction regarding what to include and exclude were sometimes fuzzy and required judgement calls.
Evaluation Criteria. Programs were included if they had been evaluated using either a randomized-trial design or a quasi-experimental design that used an adequate comparison group. Studies were required to have both pre and post-findings, and preferably follow-up data to examine the duration and stability of program effects. In addition, it was required that the programs have a written manual that specifies the model and procedures to be used in the intervention. Finally, it was necessary to clearly specify the sample and their behavioral and social characteristics.
Literature Review: Sources and Process
Given the quality-assurance inherent in the peer review process, the search primarily focused on refereed professional journals, which were searched via available databases. These databases included: PsycINFO, Social Science Abstracts, Sociological Abstracts, ContentsFirst (journal tables of contents), and Education Abstracts (ERIC).
From a search of these databases articles were identified related to a core group of programs. Government reports, meta-analyses, reviews, annotated bibliographies and relevant books and book chapters were also reviewed. Among these were reports from the Institute of Medicine, National Advisory Mental Health Council, American Psychological Association, Department of Education, National Institute of Health, and National Institute of Mental Health, as well as reviews and meta-analyses by Kazdin (1988) , Durlak & Wells (1997, 1998) , Rickel & Allen (1987) , Albee & Gulotta (1997), Weissberg & Greenberg (1998), and Catalano et al (1998), among others.
Relevant internet sources were checked such as the web pages of the Centers for Disease Control and Prevention, the Society for Prevention Research and Early Career Preventionists Network, the Collaborative for Social and Emotional Learning, the Center for the Study and Prevention of Violence, the Oregon Social Learning Center, and NIMH Prevention Research Center. These sources were cross-checked against the core group of programs to identify and secure articles for additional programs. With each new document obtained, the reference list was reviewed against the list of identified programs to further guard against omissions.
From these collective sources a set of core programs was identified for inclusion in this report. In nearly all cases, the principal investigators was contacted during the review process to address specific questions or review the information for accuracy.
The review led to the identification of over 130 programs.
Of those, 34 met the criteria discussed above and thus are included
in this report. In addition, the report discusses a number of
promising but as-of-yet, unproven models, as well as some programs
that have demonstrated effects in areas related to or often comorbid
Universal prevention programs are the broadest forms of preventive intervention in the continuum promoted by the Institute of Medicines 1994 report. Universal programs may address a group as large as the entire school-age population in a country, as is the case in the Norwegian Intervention Campaign Against Bully-Victim Problems (Olweus, 1993), or may be more narrowly directed at children in a specific grade level or a specific group identified by characteristics unrelated to risk.
In our review, we have identified fourteen universal preventive interventions which have undergone a quasi-experimental or randomized evaluation and been found to produce positive outcomes in either (a) specific symptoms of psychopathology such as aggression, depression or anxiety, or (b) commonly accepted risk factors associated with psychopathology such as impulsiveness, cognitive skill deficiencies or antisocial behavior.
Before discussing the identified programs, it is instructive to point out some of the advantages and disadvantages of universal approaches to prevention, as well as the "trade-off" between universal and targeted (indicated or selective) approaches (Offord, 1996). A potential disadvantage of universal programs is that, based on the relatively low prevalence of psychopathology among children, much of the effort will be spent on children who may not otherwise have developed mental health problems anyway (although the value of promoting competence and positive mental health cannot be overlooked). Further, because of the relatively low dosage provided by most universal interventions, they might not provide sufficient duration or intensity to alter developmental pathways of children already at significant risk for psychopathology. Offord (1996) also raises the question of whether universal programs will have the greatest impact on those at lowest risk, though the findings of some programs (Kellam, Ling, Merisca, Brown, & Ialongo, 1998; Reid, Eddy, Fetrow, & Stoolmiller, in press) contradict this theory by demonstrating stronger effects for more at-risk subgroups.
Among the advantages of universal programs is the reduced risk of the potentially deleterious effects of labeling which may be more likely in targeted interventions whose screening instruments will undoubtedly produce "false positives". Another advantage is the potential for a single preventive intervention to reduce or prevent multiple problems. A growing body of research shows that many poor outcomes such as psychopathology, substance abuse, delinquency, school failure, and teen pregnancy have overlapping associated risk factors and a significant degree of comorbidity. Because of their focus on risk reduction and health promotion, universal preventive interventions often produce reductions in multiple problem areas, as the program descriptions below will demonstrate. In addition, universal programs may also promote well being and enhance resilience.
Durlak (1995) provides another perspective. He points out that if only 8% of well-adjusted children go on to have serious adjustment problems as adults (as opposed to 30% of clinically dysfunctional children), the well adjusted children will represent 50% more of the population of maladjusted adults, based on real numbers. It may then be beneficial to provide universal preventive interventions regardless of the low prevalence rate of childhood psychopathology.
As the Institute of Medicine (1994) states, the decision to implement a universal intervention must weigh the potential benefits, given the risk of psychopathology among the target population, against the cost of implementing such an intervention for a broad (universal) population. As research continues to increase our knowledge of causal risk factors and their relative importance, this equation may balance more in favor of universal preventive interventions.
Effective Preventive Interventions: Universal Programs
Fourteen universal programs were identified as meeting our criteria for inclusion based on study design and positive outcomes related to psychopathology. For ease of discussion, they can be classified into 4 categories: violence prevention programs; more generic social/cognitive skill-building programs, programs focused on changing the school ecology, and multi-component, multi-domain programs. Although we will use this typology for discussion purposes, in actuality the programs do not fall along a linear continuum and may include characteristics of more than one of the above categories. This typology is useful however in that it is somewhat representative of the recent progress of prevention science, as the field continues to move in the direction of comprehensive, multi-system programs that target multiple risk factors across both individual and ecological domains.
The following paragraphs briefly describe the identified programs in terms of program content and evaluation results. More detailed descriptions of each program are provided in Appendix A. In the descriptions that follow, unless stated otherwise all outcomes are significant at the p<.05 level or better.
Violence prevention programs Programs that focus specifically on preventing or reducing violence, usually through curriculum-based teaching of nonviolent conflict resolution or decision-making skills, have seen mixed results. The Second Step program is a curriculum-based model that focuses specifically on skills to understand and prevent violence. Second Step aims to reduce or prevent aggression by teaching anger management, empathy and impulse control. Grossman and colleagues (Grossman, Neckerman, Koepsell, Liu, Asher, Beland, Frey, & Rivera, 1997) evaluated Second Step in a randomized controlled trial with approximately 800 primarily European-American elementary students from 12 schools in Washington State. Post-test data showed significant reductions in aggression and increases in neutral or prosocial behavior as measured by coded observations, though there were no significant effects found on parent or teacher ratings of behavior problems. Reductions in observer-rated physical aggression in the classroom were maintained at 6-month followup. The program also includes an unevaluated video-based parents guide to assist parents in reinforcing the lessons at home.
Farrell, Meyer, and White (in press) evaluated the Responding in Peaceful and Positive Ways (RIPP) program. The 25 session RIPP program focuses on social/cognitive skill-building to promote nonviolent conflict resolution and positive communication. Program activities include team building and small group work, role playing, and relaxation techniques. In a randomized trial with approximately 600 students from three middle schools in Richmond, Virginia, Farrell and colleagues report that students made significant gains on measures of decision-making knowledge and use of peer mediation, but those gains were not found on student self-reports of behavioral changes. Although significant reductions were achieved in weapon carrying (immediate post-test) and in-school suspensions (post-test and 6 month follow-up) as measured by school disciplinary data, after controlling for pretest differences and attrition no significant effects were found for fighting, out-of-school suspension, or 4 self-report measures of behavior and adjustment.
General social/emotional cognitive skill-building programs A number of the programs identified in our review focus on generic social/emotional cognitive skill-building as a means to reduce psychopathology, a wide range of deleterious outcomes which share common risk factors, as well as to promote social/emotional competence. As research, experience and practicality have dictated, these programs are often school-based and directed at elementary students.
Among the pioneers in this area are Shure and Spivack (1982), who developed the Interpersonal Cognitive Problem-Solving (ICPS) program and conducted some of the early research on the potential impact of cognitive problem solving ability on reducing poor outcomes for children. A classroom teacher generally implements the ICPS program with small groups of children. The program begins by teaching children fundamental skills related to language, thinking, and listening and progresses to practicing more complex interpersonal problem solving through dialogues and role-playing. ICPS has been implemented widely in diverse schools throughout the country and has undergone a number of evaluations. In trials with both preschool and elementary-aged populations, Shure and Spivack have demonstrated that ICPS can significantly improve cognitive problem solving abilities and reduce inhibition and impulsivity, with effects lasting through 1 year followup (Shure, 1997; Shure and Spivack, 1988). However, no data has followed children for more than one year post-intervention and there have been no findings reporting reduction in psychiatric symptoms.
Promoting Alternative THinking Strategies (PATHS) is another elementary-based program to promote social/emotional competence through cognitive skill-building. With an emphasis on teaching students to identify, understand and self-regulate their emotions, PATHS also adds components for parents and school contexts beyond the classroom to increase generalizability of the students newly-acquired skills. Greenberg and colleagues have conducted several randomized controlled trials of PATHS with a variety of populations (e.g. with regular education students, with deaf children, with behaviorally at-risk students, and as a universal intervention in a multi-component comprehensive program). In a randomized controlled trial with 200 second- and third-grade regular education students PATHS produced significant improvements in social problem solving and understanding of emotions at post-test. Compared to controls, general education intervention children show one year follow-up improvements on social problem-solving, emotional understanding, self-report of conduct problems, teacher ratings of adaptive behavior, and cognitive abilities related to social planning and impulsivity (Greenberg & Kusche, 1997, 1998a; Greenberg, Kusche, Cook, & Quamma, 1995). These improvements were maintained at 1-year followup and, more importantly, additional significant reductions in teacher and student reports of conduct problems appeared at 2-year followup.
For children with special needs, results indicated post-test improvement on teacher-rated social competence, child report of depressive symptoms, and emotional understanding and social-cognitive skills. At one-year and two-year follow-up, both teachers and children separately reported significant improvements in both internalizing (e.g., depression and somatic complaints) and externalizing behavior problems, as well as improved social planning and decreased cognitive impulsivity (Greenberg & Kusche, 1997, 1998b; Greenberg, Kusche, Cook, & Quamma, 1995).
The Improving Social Awareness Social Problem Solving (ISA-SPS) Program targets the transition to middle school as a normative life event which places children at increased risk for poor outcomes. ISA-SPS focuses on individual skill-building to promote social competence, decision-making, group participation and social awareness. Through a two-year program given to students prior to their transition to middle school, ISA-SPS seeks to bolster students resilience in the face of the many stresses related to school change.
In a quasi-experimental design with a non-equivalent control group, Elias and colleagues found improvements in youth self-report of coping with stressors related to middle school transition and teacher reports of behavior (Bruene-Butler, Hampson, Elias, Clabby, & Schuyler, 1997; Elias, Gara, Schuyler, Branden-Muller, & Sayette, 1991). More importantly, they report significant reductions in measures of adjustment and psychopathology at six-year followup: the comparison boys had higher rates of involvement with alcohol, violent behavior toward others, and self-destructive/identity problems, whereas comparison girls had higher rates of cigarette smoking, chewing tobacco, and vandalism. As an example of action research, ISA-SPS has undergone continuous testing and refinement since its inception in the early 1980s. The program, now known as Social Decision-Making and Social Problem Solving (SDS-SPS), has evolved into a more comprehensive effort with a greater ecological focus on school system change and has been expanded to address all grade levels. ISA-SPS has seen significant replication through support from the U.S. Department of Education and the William T. Grant Foundation.
Weissbergs Positive Youth Development Program (PYD) is another example of a school-based program focusing on student skill-building. The 20 session curriculum to promote general social competence and refusal skills related to alcohol and drug use was evaluated with 282 mostly African-American students from one urban and one suburban middle school in Connecticut. In a quasi-experimental study with a non-equivalent control group, Weissberg and colleagues found the program produced significant improvements in coping skills and students ability to generate alternative responses to hypothetical situations, as well as teacher reports of several measures of social adjustment including conflict resolution with peers, impulse control, and popularity (Caplan, Weissberg, Grober, Sivo, Grady, & Jacoby, 1992). Interestingly, although this program primarily targeted outcomes related to substance abuse, the program produced no significant effects on measures related to drugs, cigarettes or wine, and only marginal effects related to alcohol.
The PYD program has since been combined with an earlier 16-session version (called the Yale-New Haven Social Problem Solving program) to create the broader, 45-session Social Competence Promotion Program for Young Adolescents (SCPP-YA). Weissberg, Barton, and Shriver (1997) report that in a controlled pre-post study SCPP-YA students maintained stable levels of self-reported antisocial and delinquent behavior while control students saw a 36.8% increase. These findings however have not yet been published in a refereed journal.
Unlike the other universal preventive interventions discussed in this report, which focus primarily on externalizing behavior problems, two universal programs in Israel have demonstrated positive effects on internalizing behavior and suicidality. Klingman and Hochdorf (1993) describe a program which demonstrated positive effects on suicide risk for junior-high students in Israel. In a randomized trial with 237 8th grade students, the 12-week group cognitive-behavioral program produced significant reductions in suicidality, as measured by the culturally adapted Israeli Index of Potential Suicide (IIPS), among treatment boys. Effects for girls on the IIPS did not reach the level of significance. Likewise, Orbach and Bar-Joseph (1993) also report on a universal suicide prevention program which demonstrated a significant reduction in suicidality, in this case among 11th grade students from 6 high schools in Israel. This introspective, cathartic program was evaluated in a randomized trial examining 393 students (including some conduct disordered students) again using the IIPS. Across all schools, the authors report significant effects on suicidal tendencies, coping skills, and ego identity. Neither of these suicide prevention programs have shown effects on suicidal behavior, examined distal effects, or been replicated.
Programs focused on changing school ecology Rather than focusing primarily on the individual, ecologically-focused programs attempt to address contextual variables in the childs home or school as a means to prevent or reduce psychopathology or other negative outcomes. The School Transitional Environment Project (STEP) for example, based on the Transitional Life Events Model, focuses on changing the school ecology to be less threatening to students during the transition from elementary to middle school or from middle school to high school. STEP seeks to reduce the complexity of the new school environment, to redefine the role of the homeroom teacher as more supportive, and to create a stable support mechanism through a consistent set of peers and classmates. Through a series of evaluations and replication studies with primarily urban minority students, with study populations of as many as 2,000 students, Felner and colleagues (Felner and Adan, 1988; Felner, Ginter, & Primavera, 1982; Felner, Brand, Adan, Mulhall, Flowers, Sartain, & DuBois, 1993) found that STEPs restructuring of the school environment produced significantly lower levels of stress and reductions in anxiety, depression and delinquent behavior. In an experimental study comparing STEP with a more general intervention which taught generic coping and problem-solving skills, the STEP students experienced a better adjustment to school change, especially in academic progress (Felner, et al., 1993).
The Child Development Project (CDP) focuses primarily on changing the school ecology to create schools which are "caring communities of learners". CDP provides school staff training in the use of cooperative learning and a language arts model that fosters cooperative learning, as well as a developmental approach to discipline that promotes self-control by engaging students in classroom norm-setting and providing them with opportunities to actively participate in classroom decision-making. School-wide community-building activities are used to promote school bonding, and parent involvement activities such as interactive homework assignments reinforce the family-school partnership. The program was evaluated with approximately 4,500 third- through sixth-grade students in 24 diverse schools throughout the United States and was found to produce significant reductions in self-reported delinquent behaviors including weapon carrying, skipping school, and vehicle theft (Battistich, Schaps, Watson, & Solomon, 1996). It is important to note that effects were found only after controlling for degree of implementation (i.e. findings were only significant for "high level of implementation" schools), reaffirming the importance of fidelity in implementation.
Kellam et al. (1998) describes a randomized controlled trial with nearly 700 first grade students from 19 elementary schools in Baltimore, Maryland. The study assessed the impact of the Good Behavior Game, a team-based classroom program designed to improve childrens social adaptation to the classroom related to rules and authority, as compared to Mastery Learning, an intervention which promotes reading competency through group goal setting, and a control group. The Good Behavior Game divides the classroom into three heterogeneous teams that compete for rewards based on not exceeding established classroom standards for behavior. At post-test (end of grade one) Kellam and colleagues report significant reductions in teacher and peer ratings of aggression, as well as teacher ratings of shy behavior (a strong risk factor for negative outcomes when coupled with early aggression) among the Good Behavior Game students. Mastery Learning students showed significant improvement in reading competency. At six-year follow-up, there were no main effects of the Good Behavior Game but there was some indication that males rated as highly aggressive at first grade showed treatment effects on teacher-rated aggression. The results should be interpreted with caution based on the potential confounding and threat to internal validity of using ratings of teachers and peers directly involved in the intervention.
The Intervention Campaign Against Bully-Victim Problems (Olweus, 1993) is a nationwide program undertaken in Bergen, Norway in 1984-85 to reduce bullying and related victimization among elementary and middle school children. The program consisted of providing all teachers in Norway with a 32-page booklet that described current knowledge on the scope, cause and effects of school bullying and provided detailed suggestions for what schools and teachers could do to reduce and prevent bullying. An abbreviated 4-page folder on bullying was also provided through the schools to all families in Norway with school-age children. A 25-minute video containing vignettes of bullying situations was also made available to schools. Finally, a brief questionnaire related to bullying was administered to students. The questionnaire was considered part of the intervention as it was intended to act as a catalyst for awareness and discussion of the problem of bullying.
Olweus (1991) conducted a quasi-experimental (staggered cohort) study of the campaign with approximately 2500 students in grades 4-7 from 42 elementary and middle schools. The students were divided into 4 age/grade equivalent cohorts of 600-700 students, with roughly equal numbers of boys and girls in each. Examining data collected at 4 months pre-intervention and 8 and 20 months post-intervention Olweus reports reductions of 50% or more in bully/victim problems for boys and girls across all grades (4-9), with more marked effects after 2 years than after 1 year. Olweus also reports reductions in general antisocial behavior such as vandalism, fighting, drunkenness, theft and truancy, though the published accounts do not address the statistical significance of these findings or the validity of the measurements used.
Multi-domain, multi-component programs The Linking the Interests of Families and Teachers (LIFT) program attempts to decrease risk and increase protective factors related to future violence and delinquency. LIFT focuses on the home, the individual student, the classroom and the peer group. In the home, LIFT works to teach parents effective forms of discipline and supervision, including consistent limit-setting and parental involvement. At school, a twenty-session program is taught to increase students social and problem solving skills and help them resist negative peer groups. Finally, LIFT uses a version of the Good Behavior Game (see description above) to reduce inappropriate physical aggression on the playground.
Reid, Eddy, Fetrow & Stoolmiller (in press) conducted a randomized controlled trial with 671 children and their families from 12 public elementary schools in high-risk neighborhoods in Eugene, Oregon. At post-test, Reid and colleagues report reductions in playground aggression, with the largest effect size among the most aggressive children, as well as improvements in family problem-solving. At 30 months post-test, children from the treatment group were also significantly less likely to have been arrested.
The Seattle Social Development Project (Hawkins, Catalano, Morrison, O'Donnell, Abbott, & Day, 1992) is a comprehensive universal prevention program that addresses multiple risk and protective factors across both individual and ecological domains (individual, school, and family). With a strong emphasis on creating and maintaining strong school and family bonds, the program combines modified teacher practices and parent training across a six-year intervention period. Classroom teachers were trained in SSDP instructional methods with three major components: proactive classroom management, interactive teaching, and cooperative learning. These teaching approaches were used in combination with (a) classroom-based cognitive and social skills training in 1st (Spivak & Shures ICPS Curriculum, see above) and 6th grade (refusal and life skills); and (b) parent training that emphasized child behavior management in 1st or 2nd grade, academic support in 2nd or 3rd grade, and preventing drug use and antisocial behavior in 5th or 6th grade.
To assess the effects of full intervention and late intervention, a nonrandomized controlled trial with three conditions was created. The full intervention group received the intervention package from grade one to six. The late intervention group received the intervention package in grades five and six only, and the control group received no special intervention. 598 students were involved in the follow-up at age 18, six years after intervention. The findings indicated that students in the full intervention group reported significantly stronger attachment to school , improvement in self-reported achievement and less involvement in school misbehavior than did controls (Hawkins, Von Cleve, & Catalano, 1991; Hawkins, Catalano, Kosterman, Abbott, & Hill, in press). While no effects were shown for either the full or late intervention groups for lifetime prevalence of cigarettes, alcohol, marijuana or other illicit drug use at age 18, significantly fewer subjects in the full intervention group than in the control group had committed violent acts, reported heavy alcohol use in the past year or engaged in sexual intercourse. There were no differences between the late intervention and control conditions; this provides a strong argument for beginning social competence programs early in the elementary years and continuing them across different developmental phases.
Promising programs In addition to the universal preventive interventions described above, our review identified a number of other programs that appear promising but do not fit the criteria for inclusion in this report. These programs are not included in this review (excluded from the Appendix) because they lack a controlled design, contain very small samples, or the findings are indirectly related to mental health outcomes.
Although a number of programs include a secondary component for parents, the Effective Black Parenting Program (Myers, Alvy, Arrington, Richardson, Marigna, Huff, Main, & Newcomb, 1992) places a primary emphasis on the importance of family ecology by focusing on parental skill-building and family management. The program, developed specifically for African-American families with elementary-aged children, teaches cognitive-behavioral parenting skills with an emphasis on promoting a culturally-relevant style of child self-discipline (as opposed to a more traditional authoritative discipline style). The program showed promising results in a non-randomized control study, but the small sample size and the selective measurement of only families that did not dropout and came to more than half of the sessions limits the generalizability of the findings.
A number of social competence enhancement and violence prevention programs also show promise given that they incorporate best practices as identified in a number of recent national reports (Drug Strategies, 1998; Elias, 1997), however none of these has demonstrated effects on symptoms of psychopathology. These include the Social Skills Training Program (Rotheram, 1982), the Resolving Conflict Creatively Program (RCCP Aber, Jones, Brown, Chaudry, & Samples, 1998), the Quest Program (Laird, Syropolous, Black, & Beckley, 1996), Peacebuilders (Embry, Flannery, Vazsonyi, Powell, & Atha, 1996), the Positive Adolescent Choices Training program (PACT Hammond and Yung, 1991), Aggression Replacement Training/Skillstreaming the Adolescent (Goldstein, Sprafkin, Gershaw, & Klein, 1980; Goldstein and Glick, 1987, Goldstein, 1988), and the School Development Program (Haynes, Comer, and Hamilton-Lee, 1988).
Effective programs beyond the scope of this review A number of programs with significant effects based on well-designed evaluations do not fit within the primary focus of this review (psychopathology), but bear some discussion given the common comorbidity of psychopathology with other problems such as substance abuse. Four well-evaluated programs for the reduction of substance abuse may have the potential to reduce symptoms of psychopathology, but no such data has been reported to date. In a number of randomized controlled trials, Botvin and colleagues have shown the Life Skills Training Program to be effective in significantly reducing tobacco, alcohol, marijuana and polydrug use (Botvin, Baker, Dusenbury, Tortu, & Botvin, 1990). Evaluation results support the long-term effectiveness of the program, as well as its generalizability. Similarly, Pentz and colleagues (Pentz, Mihalic, & Grotpeter,1997) report that the Midwestern Prevention Project, in a series of quasi-experimental studies, showed significant reductions in cigarette and marijuana use. Finally, Project Northland (Perry, et al., 1996) and Project ALERT (Ellickson, Bell, & McGuigan, 1993) have also demonstrated significant effects on substance use.
Two universal parenting programs should also be mentioned. They are not included in the review because the refereed journal articles on these programs only document changes on observed parent-child interactions, and child substance use outcomes, but not child conduct problems. However, these findings are important as parental behavior, family interaction patterns, and child substance use are critical factors in the development and maintenance of conduct problems. The first program, Preparing for the Drug Free Years (PDFY; Hawkins et al., 1988; Hawkins, Catalano, & Kent, 1991), consists of 5, 2-hour, multi-media sessions designed to reduce family-related risk factors and enhance family bonding. The second program, The IOWA Strengthening Families Program (ISFP: Molgaard & Kumpfer, 1993) is an adaptation of the Strengthening Families Program originally developed by Kumpfer and colleagues (Kumpfer, DeMarsh, & Child, 1988). Both programs are currently being evaluated in several randomized and controlled studies. The samples in these studies are middle school students and their families.
Kosterman and colleagues (Kosterman, Hawkins, Spoth, Haggerty, & Zhu, 1997) found that after participating in Preparing for the Drug Free Years (PDFY; Hawkins et al., 1988; Hawkins, Catalano, & Kent, 1991), observations of parent-child interactions indicated significantly higher proactive and lower rates of negative communication, compared to controls. Spoth and colleagues (Spoth, Redmond, & Shin, 1998; Redmond, Spoth, Shin, & Lepper, in press) replicated these findings using PDFY as well as showing improvements in parents management of child behaviors. In another outcome study, positive direct effects on general child management skills were also demonstrated (Spoth, Redmond, Haggerty, & Ward, 1995). Latent transition and log-linear modeling analyses indicated that PDFY showed effects on both delayed initiation and progression of substance use (Spoth, Reyes, Redmond, & Shin, in press).
Spoth and colleagues (Spoth, Redmond, & Shin, 1998; Redmond, Spoth, Shin, & Lepper, in press) have reported similar findings for The IOWA Strengthening Families Program (ISFP: Molgaard & Kumpfer, 1993) indicating significant intervention effects on global parenting dimensions and indirect effects via intervention-targeted parenting behaviors. At one- and two-year follow-up assessments, significant intervention-control differences indicated a 60% reduction in initiation of alcohol use (Spoth, Redmond, & Lepper, in press). Similar to PDFY findings, the ISFP intervention delayed initiation of substance use at the two-year follow-up (Spoth, Reyes, Redmond, & Shin, in press).
Unpublished findings on ISFP (Spoth, personal communication, 1999) have examined intervention effects on child self-reports of problem behaviors (theft, physical aggression, vandalism, and other delinquent behaviors). At the two-year follow-up there were significant intervention-control differences on child self-reports. Earlier unpublished analyses (Spoth, personal communication, 1999) showed ISFP effects on school-related problem behaviors, and a combination of child and parent reports of affiliation with antisocial peers at the one-year and two-year follow-up assessments, though not at the post-test.
The programs referenced above have produced significant positive outcomes related to substance abuse or individual-child protective factors such as comportment or academic achievement, but either were not subject to carefully designed studies or were not evaluated in terms of their potential impact on psychopathology. This points to what can be seen as a general lack of breadth in outcome measurement for universal programs, a problem due in large part to the categorical nature of funding which promotes a view of prevention narrowly related to specific outcomes. Given the common comorbidity and overlap of associated risk factors, programs and evaluation studies should take a more global approach, measuring a broader range of outcomes.