Three diagnoses currently comprise the disruptive or externalizing behavior disorders of childhood; oppositional defiant disorder (ODD), conduct disorder (CD), and attention deficit hyperactivity disorder (ADHD) (DSM-IV; American Psychiatric Association, 1994). Although there is some similarity in these disorders, they are considered independent and unique diagnoses. Children with ADHD exhibit elevated levels of inattention and hyperactivity-impulsivity. ODD is characterized by a consistent pattern of defiant and disruptive behavior. Although ADHD children can be disruptive, they generally lack the negative quality that is the primary feature of that diagnosis. However, about 40% of children with ADHD will go one to develop significant conduct problems (Offord et al, 1992). CD includes all of the features of ODD but is a more severe and persistent diagnosis: the primary diagnostic criteria include aggression towards people and animals, destruction of property, deceitfulness or theft, and school or home rule violations. CD includes delinquent behaviors that are violations against individuals or property but it is not the same as "delinquency". This is an important distinction to make in terms of definitions of psychopathology. Many children and adolescents who engage in delinquent activity will also be conduct disordered (prevalence estimates vary but are generally between 50% and 90%) but there are some juvenile delinquents that do not have the diagnosis (Otto, Greenstein, Johnson, & Friedman, 1992).
Disruptive behavior disorders are among the most prevalent and stable child psychiatric disorders (Costello, 1989). Forty percent of children diagnosed with conduct disorder between the ages of 8 and 12 still have the disorder four years later (Offord et al., 1992). Many of the most serious and costly adult mental health outcomes and societal problems (e.g. delinquency, substance use, and antisocial personality disorder) have their origins in early conduct problems. Treatment for conduct disorder has demonstrated positive short-term outcomes but there is less evidence for long-term success (Kazdin, 1995). There are a number of reasons why treatment with younger children, or more ideally, prevention at, or prior to, symptom onset is more likely to be effective. Conduct disorder is one of the most difficult conditions to remediate because the disorder is often supported in multiple contexts, the risk factors associated with it tend to cluster together and are related in complex ways, and each risk factor tends to set the stage for increased risk in the next phase of development (Reid & Eddy, 1997).
Current State of Knowledge
Compared to other mental health disorders, a substantial amount of basic research has been conducted over the last two decades on the disruptive behavior disorders. We now have sophisticated developmental models of how these problems develop (Conduct Problems Prevention Research Group, 1992; Loeber & Dishion; Patterson, DeBaryshe, & Ramsey, 1989; Reid & Eddy, 1997) and an awareness of the risk and protective factors involved in their initiation and maintenance. There is still substantial work to be done, however, in understanding the mechanisms that link these proximal and distal factors, and how they operate over time to increase the likelihood of disorder (IOM, 1994).
Less is known about the developmental model of ADHD and the model that will be described in this section is best suited to describe the development of oppositional defiant disorder and the early-onset subtype of conduct disorder. It is important to remember that the precursors of the disruptive behavior disorders are a heterogeneous set of behaviors that are relevant to all three diagnoses. Many programs focus on these with the intent of preventing the broader groups of negative adolescent outcomes that cluster together and are associated with conduct disorder. Although all three disorders share environmental and biological-genetic components, it is likely that neurological factors that regulate activity and attention play a more substantial role in ADHD.
Research has shown that parents of children with disruptive behavior disorders tend to be less effective in managing their childrens behavior and often engage in practices that actually contribute to, and sustain, their childrens maladaptive behavior (McMahon & Wells, 1989; Patterson, 1982). High levels of coercive and punitive discipline, the use of frequent reprimands, and a lack of monitoring have all been linked to elevated levels of child aggression and delinquency (Dishion, 1990; Eron, Huesmann & Zelli, 1991; Pettit, Bates, & Dodge, 1993). Thus, parenting is both an important risk factors as well as factor that may be malleable and thus a target for preventive intervention.
Children with behavior disorders have poorer social skills, higher rates of cognitive distortions (e.g. hostile attribution bias), and cognitive deficits (e.g. poor problem solving). These social-cognitive factors not only contribute to the development of problem behavior, but are also a consequence of it. Children who are disruptive have more difficulty initiating and maintaining normative peer interactions. Over time, most children who continue to display significant amounts of aggression and negative behavior are rejected by their peers and their reputations for aggressive or inappropriate behavior reinforce their status (Dodge & Somberg, 1987). In addition, rejected children are often drawn towards more deviant peers who endorse their behavior and provide increased opportunities for antisocial behavior. Finally, difficulties managing impulses, maintaining attention, and developing positive relationships with peers and adults appear to contribute to difficulties in learning and achieving in the early years of schooling. As a result, social-cognitions, academic skills, and peer social skills are considered potential targets for preventive intervention.
There is also a growing awareness of developmental models of mental disorders outside of academia, influencing federal agencies as reflected in their publications and the programs that they are willing to support (IOM, 1994; NIMH, 1998). As an example, in the early 90s, NIMH funded a series of demonstration projects to apply the knowledge of developmental models of conduct problems to actual prevention programming.
Effective Preventive Interventions: Externalizing Programs
Many prevention programs do not have the specific goal of reducing the diagnosis of CD, but rather attempt to change levels of symptomology, or only affect the mediating mechanisms (proximal risk factors) theoretically linked to the development of the disorder. It should be noted that in all cases, preventive interventions reviewed here are either selected or indicated, and not provided universally across the population. Interventions can be categorized by their focus (child, parent, or both), intensity, and length. Ten programs met the criteria to be included in the review.
Child-focused interventions There are a number of child-focused and parent-focused conduct problem prevention programs. In general, their results are significant, but modest and tend to fade over longer periods of time. One significant drawback in many of the child-focused programs is that they have small sample sizes and have focused exclusively on boys. Some of these programs represent the earliest forms of intervention in the field. As such, they were implemented prior to the surge in prevention science that has raised the standards for program evaluation (e.g. IOM, 1994). Many of the larger-scale projects have incorporated these components as part of more comprehensive, developmentally based programs (e.g. Conduct Problems Prevention Research Group, 1992; Tremblay, Masse, Pagani, & Vitaro, 1996).
Lochman developed and refined a cognitive-behavioral school-based intervention that focuses on developing anger management skills in aggressive elementary and middle-school aged boys (Lochman, Burch, Curry, & Lampron, 1984). The Anger Coping Program consists of 18 sessions that teach affect identification, self-control, and problem-solving skills (Lochman, 1985). Children are given the opportunity to role-play and practice these skills in a small group setting and under conditions of affective arousal. Goal setting and reinforcement are incorporated to support skill acquisition. The program has shown that immediately following the intervention it lowers boys observed disruptive and aggressive behavior in the classroom, and in some cases, improves parent ratings of aggressive behavior (Lochman, 1985; Lochman et al., 1984; Lochman & Curry, 1986; Lochman, Lampron, Gemmer, Harris, & Wyckoff, 1989). In a 7-month follow-up study, children who had received the anger coping program were more on-task in their behavior compared to controls but the differences in their disruptive-aggressive behavior evident at post-test were not maintained (Lochman & Lampron, 1988). Three years after the intervention, differences in parent-ratings of aggression and observations of disruptive-aggressive behavior were not maintained although improvements in childrens on-task behavior were maintained for those who had received a six session booster the following school year (Lochman, 1992) The program had no effect of self-reported delinquency, but did have a positive effect on self-reported substance use ratings bringing the anger control subjects into the normative range on self-report ratings.
Some child-focused programs have included normative peers as prosocial models and as a way of providing the disruptive child with opportunities to practice new skills in actual social interactions. This format also counteracts the alienation experienced by many conduct problem children that makes them vulnerable to the influence of more deviant peers. Indeed, iatrogenic effects have been found in programs where antisocial youth were grouped together (Dishion, Andrews, Kavanagh, & Soberman, 1996). Most studies have shown that the conduct problem youth benefit from this group composition.
Hudley and her colleagues (Hudley, Britsch, Wakefield, Smith, Demorat, & Cho, 1998; Hudley & Graham, 1993, 1995) incorporated normative peers in a program designed to counteract attributional biases and reactive aggression in aggressive children. In her "Brainpower Program", aggressive 10 to 12 year old boys were paired with non-aggressive peers and exposed to a 12-lesson school-based intervention focusing on improving the accuracy of children's perceptions and interpretations of others' actions. Compared to a randomized control group, teacher ratings indicated that the Brainpower program was successful at reducing their aggressive behavior immediately following the intervention. At present, this program has only been evaluated on African-American boys. Although this limits the ability to generalize the findings, the use of random assignment, the inclusion of attention-only and no-treatment controls, and the fact that one of the outcome measures was from teachers that were blind to the student's status all contribute to the strength of the findings. There has been no follow-up data to date.
Prinz, Blechman, & Dumas (1994) drew upon the influence of well-adjusted peers by integrating them into groups with aggressive students. The Peer Coping Skills Training program targeted 94, 1st to 3rd grade students with high teacher-rated aggression ratings. Students were randomly assigned to either a treatment group or control. In the treatment condition, intergrated teams of children were taught prosocial-coping skills in 22 weekly 50-minute sessions. The teams progressed through different skills and levels of difficulty; new skills were not introduced until the team had demonstrated mastery of the previous skills. This format encouraged and reinforced peer support. Outcomes measured at post-test and 6 months following the intervention supported its positive effects. Children in the PCS program were rated by teachers as significantly less aggressive than controls at post-test (p<.02) and follow-up (p<.01). Significant improvements were also noted in the intervention childrens prosocial coping and teacher-rated social skills.
Aggressive children who also are rejected by their peers are at very high risk for later delinquency and violence (Coie, Lochman, Terry, & Hyman, 1992; Ollendick, Weist, Borden, & Greene, 1992). Lochman, Coie, Underwood, & Terry (1993) designed the "Social Relations Program" which consisted of 26 social skills training sessions that focused on improving the skills needed for entrance into peer groups and positive peer play. It also trained the children in social problem solving and anger management. The majority of the sessions were held individually but eight were conducted in small groups and provided the children with some time to practice the skills they were learning. The program was evaluated on a sample (n=52) of 9 to 11-year-old, African-American children. Results indicated that compared to matched controls, the aggressive-rejected children were rated as significantly less aggressive by teachers and more socially accepted by peers at post-test. The effects of the intervention were maintained at one-year follow-up. The students in the aggressive-rejected intervention group were rated by teachers as significantly less aggressive (p<.03) and more prosocial (p<.03) compared to aggressive-rejected students in the control group.
Adult support: mentoring In the last decade there has been significant community-based interest in the provision of adult support to youth to build protection against adolescent problem behavior. This has included recreation programs, after-school programs, and mentoring. At the present time there is little controlled evaluation research to indicate whether such programs can reduce psychological symptoms or protect children from mental disorders.
An exception is the Big Brother / Big Sister (BB/BS) mentoring program recently evaluated by Tierney, Grossman, & Resch (1995) with a sample of 959 youth aged 10 to 16 from 8 BB/BS agencies in geographically distinct areas. The subjects were randomly assigned to a mentor or a wait-list control condition. Based on self-report data from the participants, youth with a mentor reported that they engaged in significantly less fighting compared to controls (p<.05) and perceived their family relationships more positively. These effects were primarily due to their significance for white males though the effect on fighting approached significance for minority females. There were no significant differences between groups in terms of self-reported delinquency but treatment subjects reported that they were significantly less likely to initiate the use of drugs and alcohol (p<.05). These findings provide some promise for the effects of mentoring programs on promoting adaptive behavior in youth but, due to their reliance on self-report assessments, they should be interpreted cautiously. In addition, the constructs that were measured are not as strongly related to mental health outcomes as other symptoms or behaviors. A more general concern is that while BB/BS is one of the strongest and well-known mentoring programs, there are significant implementation problems in most mentoring programs. Among these is the recruitment and retention of mentors for sufficient periods of time to develop close, protective relations for most youth.
Multi-component programs: involvement of children and families Although child alone and parenting alone prevention models have shown limited effectiveness, a new generation of multi-component models provides the promise of greater impact. Following from developmental models of risk and protection, interventions that target multiple environments (child, school, family, neighborhood) and multiple socialization agents (parent, teachers, peers) over extended developmental periods are probably necessary to alter the developmental trajectories of children who live in high-risk environments and are already showing prodromal signs of CD (CCPRG, 1992; Reid & Eddy, 1997).
The Adolescent Transitions Program (ATP; Andrews, Soberman, & Dishion, 1995; Dishion, Andrews, Kavanagh, & Soberman, 1996; Dishion & Andrews, 1995) is a preventive intervention that targets both at-risk adolescents and their parents to prevent further escalation of problem behaviors. The program is designed to improve the self-regulation of the teens by teaching them problem solving skills. The parent component of the program attempts to improve parent management skills. In the original ATP evaluation, observations of parent-child dyads suggested that the program was successful at improving the quality of interaction in families. The impact on adolescents behavior in school was only marginal, and for one treatment condition the adolescents behavior actually worsened over time. Based on this research, Irvine and his colleagues (Irvine, Biglan, Smolkowski, Metzler, & Ary, in press) replicated ATP using only the parent component of the program. In this study, parent ratings of adolescents behavior indicated significant treatment effects.
Tremblay and his colleagues (McCord, Tremblay, Vitaro, & Desmarais-Gervais, 1994; Tremblay, Masse, Pagani, & Vitaro, 1996; Tremblay et al., 1992; Vitaro & Tremblay, 1994) combined parent training and child social skill training in the Montreal Prevention Experiment. The program targeted 166 elementary school-age boys rated above the 70th percentile on a measure of aggressive and disruptive behavior. The subjects were randomly assigned to an intervention or a placebo control condition that lasted two years. The child component consisted of group skill training sessions in which children worked with normative peers to develop more prosocial and adaptive social behavior. Parents worked with family consultants approximately twice a month for two years to learn positive discipline techniques and how to support their childs positive behavior. Initial results did not reveal many group differences although at post-test intervention students were less likely (though not significantly) to be classified as seriously maladjusted. One counterintuitive finding was that intervention subjects were rated by their parents as significantly more disruptive (p<.02) and inattentive (p<.03) at post-test. The authors attributed this finding to changes in the mothers monitoring and ability to report accurately.
Group difference began to emerge on the follow-up assessments. Intervention students were significantly more likely to be on grade level at one-year follow-up (fourth grade) compared to controls (p<.05). When the boys were 11 and 12 there were a number of significant differences between the groups. At the three-year follow-up when the boys were age 12, treatment subjects were significantly less likely than control boys to engage in fighting according to teacher report (p<.03) or to be classified as having serious adjustment difficulties. According to self-report data from age 10 to age 12, treatment boys were also significantly less likely to engage in delinquent activity compared to controls. At age 12, peer nominations of aggression from the best friends of boys in the treatment group were significantly lower than those of the control groups best friends (p<.05). Effects of the treatment on other forms of antisocial behavior (e.g. self-reported stealing) and substance use continued into early adolescence (age 11 to 15). The results of the Montreal Prevention Experiment reflect the importance of extending assessments beyond the post-test point particularly when the behaviors being targeted by the intervention or more likely to occur later in development (e.g. delinquency). In this program, group differences between the intervention and control group were apparent in multiple domains (i.e. academic, social, behavioral), emerged over time, and became increasingly significant.
Another multi-component program that combines parent and child-focused interventions is the First Steps Program (Walker, Kavanagh, Stiller, Golly, Severson, & Feil, 1998; Walker, Stiller, Severson, Feil, & Golly, 1998). This program intervenes with children and teaches them more adaptive behavior that is likely to foster social and academic success. The initial phase consists of a comprehensive screening process (Early Screening Project) which identifies kindergarten children exhibiting elevated levels of antisocial behavior. Families with an at-risk child receive a 6-week home intervention in which program consultants help them develop ways of supporting their childs adaptive behavior. In school, target children participate in a classroom-based, skill-building and reinforcement program that lasts two months. The program was evaluated with 42 subjects (two cohorts) using a randomized, experimental design. Teachers in this study reported significantly less aggressive (p<.001) and maladaptive (p<.001) behavior for intervention students compared to those in the control group at postintervention. Immediately following the intervention, the teachers also rated the intervention students as significantly more adaptive than the controls (p<.001) and observations indicated that program students showed more time engaged in academic activity (p<.05). There were no group differences on teacher ratings of withdrawn behavior. Cohort I students were assessed again in first and second grade. Cohort 2 students were followed into first grade. Treatment effects were maintained for both groups at these time points. The authors conducted a replication of the First Steps program (Golly, et al., 1998) with a new sample of twenty kindegarten students. The postintervention results were almost identical to those found in the original trial. A number of independent replications are currently being conducted, but outcome data were not available at the time of this report.
The Earlscourt Social Skills Group Program (Pepler, King, & Byrd, 1991; Pepler, King, Craig, Byrd, & Bream, 1995) is a multi-component program that targets three domains: the child, the parents, and the classroom. Students age 6 to 12 exhibiting aggressive and disruptive behavior (according to both teacher ratings and principal reports) are eligible to participate in the program. The primary intervention is social skill training provided in small groups, twice weekly over the course of 12 to 15 weeks. Training sessions are offered to parents but not required. Classroom presentations, teacher involvement, and homework assignments are all utilized to generalize the skills to the classroom setting. The evaluation included 74 boys and girls who were randomly assigned to the intervention or a wait-list control group. Findings revealed that teachers rated intervention students as exhibiting significantly less externalizing behavior than controls at post-test (p<.05) and that these were clinically significant changes in symptomatology (defined by the authors as an improvement of .5 SD). It is important to note that significant group differences were only found on teacher ratings and parents failed to see significant behavior changes in the intervention children.
Recently a consortium of prevention researchers have developed Fast Track, a school-wide program that integrates universal, selective, and indicated models of prevention. It is intended to provide a comprehensive longitudinal model for the prevention of conduct disorders and associated adolescent problem behaviors (Conduct Problems Prevention Research Group, 1992). This randomized clinical trial involves 50 elementary schools in four U.S. urban and rural locations. The universal intervention includes teacher consultation in the use of a series of grade level versions of the PATHS Curriculum throughout the elementary years. The targeted intervention package includes a series of interventions that involve the family (e.g., home visiting, parenting skills, case management), the child (e.g., academic tutoring, social skills training), the school, the peer group, and the community. Targeted children and families consist of those who are identified by a multi-stage screening for externalizing behavior problems during kindergarten and they consist of the ten percent of children with the most extreme behavior problems in schools in neighborhoods with high crime and poverty rates.
Results of the first three years indicate there are significant reductions in special education referrals and aggression both at home and at school for the targeted children (Conduct Problems Prevention Research Group, 1998, 1999a, 1999b). Fast Track is predicated on a long-term model (i.e., the intervention will continue through middle-school) that assumes that prevention of antisocial behavior will be achieve by building competencies and protective factors in the child, family, school, and community. The initial results provide evidence for improved social and academic development. Results of the universal component (The PATHS Curriculum, see Universal Prevention section of this report ) at the end of grade 1 show lower students' sociometric reports of peer aggression, and improved observers' ratings of the classroom atmosphere in the intervention sample (Conduct Problems Prevention Research Group, 1999b).
Promising programs There were a number of programs that were 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. They are still considered promising because they incorporate a number of best practices. One example is the recently developed Coping Power Program, a modification of the Anger Coping Program (Lochman, 1998; Lochman, in press). The program includes a 33-session child group component and a 16-session parent group component. Initial findings indicate that the program has led to reductions in parent and teacher rated aggression at posttest and at one-year follow up (Lochman, 1998; Lochman, in press).
The FAST Program (McDonald et al., 1997; McDonald & Sayger, 1998), is a family-based preventive intervention designed to improve the protective factors in families in which children are exhibiting behavioral and academic problems. The program uses a combination of parent, parent-child, and multi-family sessions to build social support surrounding each family. FAST has only been evaluated in a non-experimental design but showed promise for its ability to build protective factors and improve overall functioning in families but it needs to be evaluated in clinical trials. The author notes that there are currently five evaluations being conducted on the program and all utilize a randomized experimental design (McDonald, 1999, personal communication).
The Contingencies for Learning Academic and Social Skills (CLASS; Hops, Beickel, & Walker, 1976) program is a child-focused program that aims to increase appropriate behavior of acting-out students by training teachers in a number of behavioral principals. The original program was developed and evaluated in the 1970s, and its most recent version is included in the First Steps to Success Program (Walker, Kavanagh, et al., 1998; Walker, Stiller, et al., 1998). As a freestanding program, CLASS is mentioned as a promising program. It has been shown to increase the positive behavior of disruptive students (Hops et al., 1978; Walker, Retana, & Dersten, 1988). There were limitations to the early evaluations in that the sample sizes were small and there were no long-term follow-up assessments in either of two studies.
Effective programs beyond the scope of this review Although efforts to prevent conduct problems have often focused on children, parents of conduct problem children have also been the targets for change. During the last 25 years there have been numerous demonstrations of the short-term effectiveness of social-learning based parent-training and education programs for families in which children are showing clinical and near-clinical levels of aggression and disruption. These were pioneered by Patterson and his colleagues (Patterson, 1982; Patterson, Chamberlain, & Reid, 1982; Patterson, Reid, Jones, & Conger, 1975) and Forehand and colleagues (Forehand and McMahon, 1981). Recent findings by Webster-Stratton using parent training, and the combination of parent training and child skills training, replicate and bolster these early efforts and demonstrate longer-term effects (Webster-Stratton, 1990; Webster-Stratton & Hammond, 1997; Webster-Stratton, Hollingsworth, & Kolpacoff, 1988, 1989). The Parent and Children Training Series (Webster-Stratton, 1992a, 1992b) has also been independently replicated in a community mental health setting and shown to produce similar results (Taylor, Schmidt, Pepler, & Hodgins, 1998).
While most of these programs were designed as treatment programs and were evaluated on clinical populations, the best practices and techniques used in these parenting programs have been more recently applied in a prevention framework in universal models as well as to selected and indicated populations. For example, in the Fast Track program the parenting training component is based on both the Parent and Children Training Series (Webster-Stratton, 1992a, 1992b), now referred to as The Incredible Years: Parents, Teachers, and Children Series, and the Helping the Noncompliant Child curriculum (Forehand and McMahon, 1981). The interface of prevention and treatment is an important issue for the field of prevention science. Some researchers would argue that interventions that improve the behavior of young children exhibiting clinical levels of behavior problems (who may or may not be diagnosed as having Oppositional Defiant Disorder) should be considered prevention programs because ODD is often a developmental precursor for Conduct Disorder. If this is the case then much of the early work in the field of parenting training would be considered "prevention". This report however adopts a more conservative definition of prevention when determining the criteria for effective programs, although the authors recognize the importance of carefully considering definitions for the field given their potential impact on program funding and dissemination.
Webster-Stratton recently adapted her videotape series (this version is referred to as "PARTNERS") to a younger, Head Start population (Webster-Stratton, 1998). This program is mentioned because it meets the criteria for an effective prevention program but was not included in the review because the age of the participants placed it beyond the scope of this review. In the evaluation study, parents of children in nine Head Start centers were randomly assigned to receive the intervention or serve as a control by only receiving the regular Head Start services. The intervention lasted 8-9 weeks and consisted of parent training groups and a teacher-training program. Results at post-test indicated significant improvements in the intervention parents on self-reported measures of discipline practices and other aspects of parenting behavior. Teacher reports indicated that parental involvement in school was higher for intervention parents compared to controls (p<.01). Observations of intervention children revealed a significant reduction in conduct problems for the intervention children compared to controls (p<.001) and teacher reports indicated significant improvements in school-based behavior (p<.05). Follow-up assessments were conducted at 12-18 months following the intervention and significant differences were maintained between the groups on measures of maternal discipline practices (p<.001), observations of maternal behavior (p<.01), and teacher reports of childrens behavior (p<.05). These results with a preschool population have not yet been replicated in an older group.
Another group of effective programs that were excluded because they were beyond the scope of the review are family therapy-based interventions that have shown significant success in treating violent and chronic juvenile offenders, and drug-abusing adolescents. These programs include Multisystemic Therapy (MST; Henggeler, Melton, Brondino, Scherer & Hanley, 1997), Functional Family Therapy (Alexander, & Parsons, 1973: Barton, Alexander, Waldron, Turner, & Warburton, 1985), and Structural Family Therapy (Santisteban, Szapocznik, Perez-Vidal, Kurtines, Murray, & LaPierre, 1996; Szapocznik, Perez-Vidal, et al., 1988; Szapocznik, Rio, et al., 1989). In general, these programs have not yet been applied to a less severe population with a prevention focus. Initial efforts are underway to adapt Multidimentional Family Therapy (MDFT; Liddle,1991) to a family-based prevention program (Liddle & Hogue, in press).
Although there is some evidence of the effectiveness of parent education and training for promoting more positive parenting, there is insufficient evidence at this time that parenting intervention alone has led to significantly reduced levels of symptomology in school-age populations over extended periods of time.