Comorbid mental disorders and substance use disorders: epidemiology, prevention and treatment

Windows of opportunity: outcome studies in the prevention of internalising disorders and externalising disorders

Page last updated: 2003

From a developmental perspective, there are likely to be optimum times and optimum methods for taking preventive action, an area that will eventually become clearer as further prevention studies are evaluated longitudinally. At this stage, although prevention has been receiving increasing press in the literature, the number of controlled, longitudinal studies is decidedly small. However, there are programs available that could be implemented and evaluated to test the hypothesised aetiological developmental trajectory of anxiety disorders, through depressive disorders, to substance use disorders, and the trajectory to substance use disorders through childhood externalising behaviour disorders. These programs are outlined below within a developmental framework, beginning with early childhood and moving to middle childhood and adolescence. Due to the suggested presence of anxiety disorders before depressive disorders in the case of internalising disorders, many of the programs presented for internalising disorders in young children focus on prevention of anxiety problems, before programs for the prevention of depression are introduced within a slightly older age group.

Early childhood, internalising disorders
Early childhood, externalising disorders
Middle childhood, internalising disorders
Middle childhood, externalising disorders
Adolescence, internalising disorders
Adolescence, externalising disorders
Additional evidence for prevention of substance use disorders by intervening with internalising disorders and externalising disorders

Early childhood, internalising disorders

In the realm of family and temperament risk factors, infancy and early childhood (children up to 4 to 5 years of age) are ideal points of prevention. One of the obstacles to determining the effectiveness of preventive efforts for children of this age is the lack of established assessment criteria suitable for use with such young children at the community level. Additionally, many of the cognitive-restructuring aspects of reducing anxiety are beyond the cognitive capacities of children in this age group, and adult modeling and shaping is the primary avenue of protection. Thus, for infants and preschoolers, the best treatment approach is working with parents (Bernstein, Borchardt, & Perwien, 1996). Knowledge of developmental needs, including differences in temperament, parental support, fostering secure attachment, and parental acquisition/modelling of coping strategies, are broad areas of prevention. These strategies provide opportunities for parents to learn patterns of interaction that support children's wellbeing, as well as skills to manage parental stress.

The most common forms of internalising disorders in this age group are anxiety problems such as Specific Phobias and Separation Anxiety Disorder. There is a body of literature showing that brief cognitive-behavioural treatments implemented through the parents are successful in reducing these problems (for a review see Dadds, Barrett, & Cobham, 1998), and in a general developmental sense, these thus offer potential as preventive interventions for substance use disorders. However, the evidence for the use of primary, secondary, indicated or selected interventions for internalising disorders in this age group is scarce.

LaFreniere and Capuano (1997) implemented a 6-month intensive home-based indicated prevention program for mothers and preschoolers. This project offered information on child development, including booklets on Development, Behaviour, Security, The Body, and Parental Needs. Additional sessions were provided to address core skills in parenting, as well as any additional personal or parental concerns presented. The aim of these sessions was to alleviate stress within the parent-child relationship. Finally, parents were assisted to build a social support network. At the conclusion of the program, anxious withdrawn preschoolers identified through teacher assessments showed significant gains in social competence, although reductions in anxious-withdrawn behaviour only approached significance. Parenting stress in the intervention group did not show a significant reduction relative to controls, although a subjective positive bias was noted in mothers who participated in the intervention.Top of page

A parent-teacher universal prevention program for children aged 4 to 5 years, aimed at reducing the incidence of internalising disorders later in childhood, was recently evaluated in Brisbane, Australia (Roth & Dadds, submitted). The project was a large-scale community program that attempted to identify children at risk in this young age group, and determine the short- and long-term effects of a prevention program through a controlled trial. Entitled, REACH for Resilience, the program aims to teach parents and teachers strategies and ways of thinking that can increase children's ability to cope with challenges, especially through adult modeling of these strategies and encouragement of children's efforts. Analysis of recruitment and retention patterns showed that, in the intervention group, the most stressed parents agreed to participate and attended the treatment sessions. In the comparison group, the most stressed parents self-selected out. At post-treatment and follow-up, the groups were not different on any of the parent and child adjustment or diagnostic measures. Thus, while the results are encouraging in terms of reaching the most needy parents, this confounds results and makes conclusions about intervention effects dubious.

Summary: At this stage, the empirical evidence is inconclusive regarding optimal prevention of anxiety disorders in early childhood. Firstly, it would be drawing a very long bow to argue at this stage that such interventions could potentially reduce incidence of substance use disorders in later life. However, drawing from the literature on resilience (Cowen, Wyman, & Work, 1996; Cowen et al., 1997), the experience of a positive and continuing relationship with a caregiver seems to be a major factor influencing resilient versus non-resilient children (Werner, 1993). Secondly, children's temperament (easily soothed, low emotionality, sociable) tends to elicit positive responses from adults as well as children, thereby assisting with the development of social competence (Fox & Calkins, 1993). Thirdly, an internal locus of control (having a sense of influence over life's events) was more evident in resilient children, and can be supported by age appropriate problem solving strategies (Shure, 1997; Wyman, Cowen, Work, & Kerley, 1993). Fourthly, an optimistic outlook predicted socio-emotional adjustment and a stronger internal locus of control (Wyman et al., 1993). Thus, prevention initiatives in early childhood might focus on developing secure attachments; modeling of appropriate coping strategies such as optimism, problem solving, and seeking social support; and ultimately taking action. Longitudinal studies are necessary to a) develop efficacious and effective programs; b) discover the specific factors necessary and sufficient to prevent the onset of anxiety disorders and build resilience; and c) track the effectiveness of these strategies over time.

Early childhood, externalising disorders

The externalising disorders that most commonly appear within early childhood are generally characterised by disruptive behaviour in the home and preschool (Tremblay, Pagani-Kurtz, Masse, Vitaro, & et al., 1995). Research shows that disruptive behaviour in early childhood represents a salient risk factor for the continued expression of behavioural disorders (Dadds, 1997; Hawkins et al., 1992;Tremblay et al., 1995) and substance use disorders (Dadds, 1997; Hawkins et al., 1992). Research further suggests that brief behavioural treatments implemented with multiple points of focus, for instance via parenting skills in the home and via social skills training in schools, can prove more effective than programs which only target one of these domains (Kazdin, 1993; Kazdin 1995). Similarly, it is recognised that programs seeking to change behaviour produced within a particular developmental context must address all of the components of that context. Thus, preventive interventions that include both parent- and child-focused components would be expected to be more optimally effective than programs that incorporate only one of these intervention targets (Coie & Jacobs, 1993; Dodge, 1993).

Considerable work has been done on the development and evaluation of tertiary treatments for externalising disorders. The most successful are parent training and family interventions, and for older children, individual or group social-cognitive work with the child. Research has supported the efficacy of behavioural family interventions in the short term and over follow-up periods of years after the termination of treatment (Miller & Prinz, 1990). The last few decades have witnessed continuous refinement of the behavioural family intervention approach. Empirical evidence and clinical experience suggests that not all parents or families benefit to the same extent from treatment (Miller & Prinz, 1990), and difficulties are commonly encountered when there are concurrent family problems, parental psychopathology, and economic hardship. Several authors have made various proposals to improve the outcome of treatment by expanding the focus of treatment to the multiple systems that provide the context for family life (Henggeler, Melton, Brondino, Scherer, & Hanley, 1997; Miller & Prinz, 1990). Of particular interest to early intervention is the Triple P approach (Sanders, 1999). The Triple P framework offers various levels of intervention intensity, from simple provision of information through to a full multisystemic, individually tailored intervention. Of the different approaches encompassed by behavioural family intervention, parent training for the treatment of younger Oppositional Defiant Disorder children has the most accumulated evidence regarding its therapeutic effectiveness. There is less evidence to suggest that behavioural family intervention is effective in altering the course of the more severe end conduct problem children, especially beyond the years of early childhood.Top of page

One example of an effective multi-focused preventive intervention administered to a select sample comes from Tremblay et al., (1995). In this study, disruptive kindergarten boys were randomly allocated to a dual focused preventive intervention condition or to a control condition. All the participants were from inner city low socio-economic neighbourhoods. The components of the dual focused intervention program included home-based parent training in effective child rearing practice, and appropriate social skills training for child participants. The child participants were compared with controls at four time points — at the end of the program, prior to puberty, at puberty, and during adolescence. Based on these comparisons the program was judged successful for the following reasons — a) compared with children in the control group, a significantly larger number of boys who undertook the intervention remained in regular and age appropriate classrooms until the end of elementary school, and b) the boys who participated in the treatment program showed significantly less delinquent behaviour at the post-intervention assessments carried out each year while the children were between 10 and 15 years of age.

Middle childhood, internalising disorders

Middle childhood appears to be an especially advantageous time for anxiety prevention and early intervention. Developmentally, this is the time when most anxiety disorders emerge, and these have been shown to be predictive of adolescent depression (Cole et al., 1998). As children's cognitive abilities mature, cognitive restructuring techniques are able to be utilised in helping at-risk children change the meaning of aversive events and experiences. This is especially important because the impact of stressful events on children appears to be largely mediated by the children's evaluation of the event in relation to their wellbeing. Dadds et al., (1998) suggest that intervention with parents is especially important with younger age groups of children, whereas for older children cognitive work and exposure may be sufficient. A further advantage for this age group is that the children can complete self-report measures, providing additional reliable and valid assessment information. It should be noted, however, that collecting assessment information from multiple sources is still vital due to the tendency of anxious children to portray themselves in socially desirable ways. Using teacher nominations in conjunction with children's self-reports seems most efficacious as each of these methods tap different types of anxiety problems, while at the same time being supported as valid assessment tools by structured interviews (Dadds et al., 1998).

Controlled clinical trials with children diagnosed with anxiety disorders have only been reported recently. The programs have included individual cognitive work to reduce threat appraisal, exposure, and enhancement of parental communication and child-rearing skills. The results are impressive with improvement maintained in 60% to 90% of cases overall in the controlled trials. Although these studies were treatment and not prevention studies, they are worth considering in some detail due to their important implications for the design and implementation of anxiety prevention and early intervention.

Two controlled treatment studies for children with a primary anxiety disorder diagnosis were conducted by Kendall and his colleagues (Kendall, 1994; Kendall, Flannery-Schroeder, Panichelli-Mindel, Southam-Gerow, & et al., 1997). These trials consisted of 16 to 20 cognitive-behaviour therapy (CBT) sessions for anxious children. The first eight weeks of the treatment involved psychoeducation regarding anxiety, and teaching children cognitive and behavioural strategies for managing and reducing their anxiety. The second 8 weeks involved practising the anxiety management skills learned previously during both imaginal and in vivo exposure to threat related situations.

In the first controlled trial (N=47), over 60% of the treatment group no longer met criteria for an anxiety disorder at post-treatment, and these gains were maintained at one-year follow-up. The second randomised clinical trial (N=94), which used the same CBT treatment as the first study, yielded very similar results. Over 50% of children diagnosed with a DSM-IV anxiety disorder pre-treatment no longer retained their diagnosis post-treatment, compared to only 6% (n=2) in the waitlisted group. For those children who did retain diagnoses at post-treatment, significant reductions were still seen in the severity of their problems. Effects were not modified by comorbidity, gender, or ethnicity. Participants completed assessment measures at eight weeks following the completion of the psychoeducation component of the treatment, allowing an examination of the effectiveness of the two different components of the treatment (psychoeducation and active exposure to anxiety provoking stimuli). Results suggested that the psychoeducation component alone was not sufficient to reduce children's anxiety disorders. However, when followed by eight sessions of active exposure, the two components together created significant reductions in diagnosable anxiety problems.Top of page

A similar treatment program (12 sessions) which involved parents as well as children was found to be superior to one which involved only children (Barrett, Dadds, & Rapee, 1996). Children (N=79) aged seven to 14 years who met criteria for separation anxiety, overanxious disorder, or social phobia were randomly assigned to one of three treatment groups — cognitive-behavioural therapy (CBT) (child only), CBT plus family/parent management, or a waitlist group. In the child plus parent treatment group, 84% of children no longer met criteria for an anxiety disorder at post-treatment, and this increased to 95% at 12 month follow-up. In the child-only treatment group, 57% of children were assessed as having no anxiety disorder at post-treatment, increasing to 70% at 12 month follow-up.

Barrett (1998) showed that similar success rates could be achieved by presenting the combined CBT-family treatment in a group format to anxious children and their parents, thereby significantly reducing costs of intervention. Barrett et al., (2001) showed durable treatment effects up to six years following treatment. Mendlowitz et al. (1999) also examined the effect of parental involvement in CBT group intervention on anxiety, depression, and coping strategies in school-age children. Similar to Barrett et al., (1996), all treatment groups showed positive change, and concurrent parental involvement enhanced the treatment effects. Cobham et al, (1998) used the same group intervention to assess the role of parental anxiety in treatment outcome, and the extent to which the second component of Barrett et al's family treatment (parent skills for managing their own anxiety) could alleviate putative poorer treatment outcomes associated with high parental anxiety. Results indicated that high parental anxiety was a risk factor for poorer treatment outcomes for anxious children, and that specifically targeting parental anxiety for intervention could overcome this risk factor in the context of a cognitive-behavioural program for the child.

Silverman et al., (1999) used a randomised clinical trial to evaluate the therapeutic efficacy of group CBT therapy versus a wait list control condition to treat anxiety disorders in children. Results indicated that group CBT, with concurrent parent sessions, was highly efficacious in producing and maintaining treatment gains. Children in group CBT showed substantial improvement on all the main outcome measures, and these gains were maintained at three, six, and 12 month follow-ups. Silverman et al. (1999) evaluated the relative efficacy of an exposure-based contingency management treatment condition and an exposure-based cognitive self-control treatment condition relative to an education support control condition for treating children with phobic disorders. 81 children and their parents completed a 10 week treatment program in which children and parents were seen in separate treatment sessions with the therapist, followed by a brief conjoint meeting. Children in both the contingency management and self-control conditions showed substantial improvement on all of the outcome measures. These gains were maintained at 3, 6, and 12 month follow-ups. Interestingly, children in the education support control condition also showed comparable improvements at post-treatment and at 3, 6, and 12 month follow-ups.

A selected prevention project targeted children (N=1786) aged seven to 14 in Brisbane, Australia (Dadds, Spence, Holland, Barrett, & Laurens, 1997). Those included in the project ranged from children who were exhibiting mild anxious features, but remained disorder free, to those who were in the less severe range of a DSM-IV anxiety disorder. An intensive screening process incorporated parent, child and teacher reports, telephone calls and face-to-face interviews. Children with a) disruptive behaviours (impulsive, aggressive, hyperactive, non-compliant), b) lack of English as a first language in the home, c) developmental delay or other problem, d) no anxiety problem according to teacher reports, and e) invalid child reports (ticked 'yes' to all items) were excluded from the sample. The final sample consisted of 128 children. Any child with severe symptoms or whose parents requested individual help for their child's anxiety were referred for individual treatment and were no longer included in follow-up assessments.

The intervention was based upon an adaptation of Kendall's Coping Cat Workbook, a 10 session program presented in group format for teaching children strategies to cope with anxiety. The sessions were conducted weekly for one hour at the child's school, in groups of five to 12 children. In addition, parents periodically attended three sessions covering: a) child management skills, b) modelling and encouraging the strategies children were learning through the Coping Koala Prevention Program, and c) how to use Kendall's FEAR plan to manage their own anxiety. The monitoring group received no intervention, but was contacted at planned intervals for follow-up assessments.Top of page

Interestingly, at post-intervention no significant differences were found between the monitoring and the intervention groups. Yet, at 6 months follow-up, the intervention group showed a significant reduction in the onset of disorder (16% onset), relative to the monitored group (54% onset). Most importantly, the success of their program in reducing the existing rate of anxiety disorder and preventing the onset of new anxiety disorders was successfully maintained at a two year follow-up (Dadds et al., 1999). These results are very promising. Given that over half of the at-risk children in the monitoring group progressed from mild anxious symptoms into a full-blown anxiety disorder, middle childhood and early adolescence appear to provide an important 'window of opportunity' for prevention initiatives.

When conducting an indicated prevention, such as described above, an important ethical caveat surrounds the potential to negatively label children who are deemed 'at-risk', and thus raise concern in parents as well as stigmatising children amongst their school peers. The Queensland project surmounted this dilemma by describing the intervention as 'a positive skill building experience', and the monitoring group provided 'an information gathering/learning exercise for researchers'. This ethical consideration should be addressed in any future programs that are designed and delivered to investigate the potential of early intervention for anxiety disorders in middle childhood to reduce later incidence of substance use disorders.

Finally, there is evidence that programs that build social skills in primary school children, without necessarily focusing on internalising disorders, can reduce the symptoms of these disorders. Such effects have been shown in the PATHS program, for example, using a range of well-designed studies with unselected, deaf, and behaviourally at-risk students (Greenberg, Zins, Elias, & Weissberg, in press).

Summary: The above review would suggest that successful prevention, early intervention and treatment in middle childhood has been achieved with regard to anxiety disorders and symptoms. Studies have been able to demonstrate long-term improvements for children up to two years post-intervention. The long-term success of these interventions has clear implications for a concomitant reduction in community costs and family distress. None of the above studies took measures of substance misuse at follow-up. However, it is reasonable to speculate that these interventions have some potential for reducing the incidence of depression and substance use disorders in the adolescent years.

Middle childhood, externalising disorders

The evidence with regard to the treatment and prevention of externalising disorders in middle childhood is also strong. There are a number of prevention programs that aim to reduce aggression and promote social skills in children via universal curriculum-based programs in schools. These may have some impact on externalising disorders but are outside the scope of this review (see Greenberg, Domitrovich, & Bumbarger, 2000). Greenberg et al. (2000) located 10 early intervention programs that have shown success in reducing externalising disorders or their risk factors. Similar to tertiary models, the majority of these utilise child-cognitive skills training, parent training, or both. Only the most recent and well-evaluated will be reviewed here.

As an example of a child-focussed program, Lochman et al. (1993) evaluated a 26 session social skills training program focusing on peer-relations, problem solving, and anger management, with a sample (n=52) of 9 to 11-year-old aggressive-rejected children. Compared to controls, the program children were rated as significantly less aggressive by teachers and more socially accepted by peers at post-treatment and at one-year follow-up. By contrast, in Lochman's (1985) program, children who had received an anger coping program were, three years after the intervention, not different from controls in terms of parent-ratings of aggression and observations of disruptive-aggressive behaviour, or in terms of self-reported delinquency. In another child-focused intervention, Tierney et al. (Big Brother/Big Sister Program: 1995) randomly assigned 959 between 10 and 16 year old adolescents to a mentor or a wait list control condition. Those with a mentor reported that they engaged in significantly less fighting, were less likely to initiate the use of drugs and alcohol, and perceived their family relationships more positively. However, there were no significant differences between groups in terms of self-reported delinquency. While encouraging, these data are based solely on self-report.Top of page

One problem with the use of group interventions for indicated externalising disorder youth is that iatrogenic effects have been found in programs where antisocial youth were grouped together (Dishion, Andrews, Kavanagh, & Soberman, 1996). In contrast, studies have found that externalising disordered youth benefit from being in groups with non-problem children. For example, Hudley and Graham (1993; Hudley and Graham, 1995) paired aggressive 10 to 12 year old boys with non-aggressive peers in a 12 lesson school-based intervention that focused on improving the accuracy of children's perceptions and interpretations of others' actions. Compared to controls, teacher ratings indicated that the program successfully reduced aggressive behaviour immediately following the intervention. There has been no follow-up data to date. A similar 22 session integration program by Prinz, Blechman and Dumas (1994) was evaluated up to six months following the intervention. Children in the program were rated by teachers as significantly less aggressive than controls at post-test and follow-up. Compared to controls, the intervention group also showed significant improvements in pro-social coping and teacher-rated social skills.

Overall, the evidence is not strong that child-focused early interventions are effective with externalising disorders. In general, their results are modest and not durable, the sample sizes are small, and due to the nature of the interventions, they are limited to older children and adolescents. The limited applicability of child-focused interventions for externalising disorders is not surprising given the literature reviewed earlier regarding the importance of early contextual factors in the development and maintenance of such disorders. More comprehensive programs that contain interventions to change problematic early parenting/and environmental issues; and others that include parenting interventions in combination with the child-focused interventions, are showing more impressive results. These are reviewed next.

Parent focused interventions generally have produced more clinically significant outcomes. As noted earlier, there have been numerous demonstrations of the effectiveness of social-learning based parent-training programs for families of children with externalising disorders. Numerous independent replications in community settings have produced significant results (Sanders, 1999). While most of these programs were designed as tertiary treatments and have been evaluated on clinical populations, a number of authors have argued that they are excellent early intervention strategies in that they effectively reduce externalising disorders early on in their developmental trajectory (e.g., Sanders, 1999). However, as we saw earlier, one limitation of a referral-based approach is that it leaves initiatives for intervention in the hands of parents, who may not seek help even in extreme situations.

Parent interventions have also been recently applied in both universal prevention and early intervention formats. Webster-Stratton (1998) has used a parent training model with young Head Start children. Because the entry procedure was based on screening of children rather than parent-referrals, the program can be regarded as a selected program. Parents of Head Start children were randomly assigned to receive the intervention or serve as a control by only receiving the usual services. The 9-week intervention consisted of parent training groups and a teacher-training program. Results at post-test and 12 to 18 months follow-up indicated significant improvements in parent behaviour, parental involvement in school, child conduct problems, and school-based behaviour.

A number of early intervention programs have been evaluated that adopt developmental models of externalising disorders and, as such, utilise multiple interventions across settings and time. This is consistent with a general view that a more comprehensive approach is necessary to alter the developmental trajectories of children who live in high-risk environments and show early signs of these disorders (Conduct Problems Prevention Research Group, 1992; Reid & Anderson, 1997).

One recent study entitled the LIFT (Linking the Interests of Families and Teachers) Intervention examined the efficacy of a universal preventive intervention in the reduction of conduct problems (Reid, Eddy, Fetrow, & Stoolmiller, 1999). LIFT was 10 weeks in duration, and targeted three distinct domains that had been identified by a developmental model of the trajectory of conduct problems. A sample of 671 first and fifth graders and their families was drawn from 12 elementary schools. The intervention condition consisted of a parent training component in the behavioural family intervention tradition, together with a playground behavioural program and a teacher parent communication program. It was hypothesised that the intervention would have significant effects on three specific areas, levels of child physical aggression in the playground, mother aversive behaviour that was displayed during interactions with their children, and teacher ratings of child peer positive behaviour over the year following the intervention. The results indicated that the intervention had significant results on child physical aggression in the playground, and on mother aversive behaviour in mother-child interactions. In addition, the results for the children's behaviour in the classroom were in the expected direction. All results were immediate and applied to both first and fifth grade participants.Top of page

Kazdin and Wassell (2000) evaluated a preventive intervention involving cognitive problem solving skills training (PSST) for the child and child/parent management training (PMT). PSST involved seeing children individually for 20 to 25 sessions to teach adaptive problem-solving skills for use in interpersonal situations such as those with family, peers, siblings, and teachers. The PMT condition was in the tradition of behavioural family intervention. For children attending school, school-based issues were included in treatment through contact with school teachers, and incorporating home-based reinforcement interventions for the school issues. In general, the children (aged between two and 14 years), their parents, and their families all responded to treatment. Children's functioning, as well as parent and family functioning, improved over the course of the intervention. This improvement was demonstrated within a range of child behavioural symptoms, parental symptoms and levels of stress, and family functioning, relationships and support (Kazdin & Wassell, 2000). Generally, larger effects were demonstrated on childrens' outcome measures, and effects of less magnitude were demonstrated on parent and family outcome measures. While the authors note that the children in this study were all under referral for conduct problems, and that similar experimental results have not been demonstrated for populations exhibiting internalising disorders, support for the generalisability of therapy based on demonstration of risk factors is warranted. For instance, it is noteworthy that improvements in both parental functioning and stress, as well as family functioning, relationships and support, have been demonstrated as important for children with both internalising and externalising disorders (Cobham et al., 1998; Kazdin & Wassell, 2000). In addition, changes in family and parent functioning may be expected to contribute to beneficial outcomes as far as long-term treatment effects for children are concerned.

In the Montreal Prevention Experiment,Tremblay and colleagues (McCord, Tremblay, Vitaro, & Desmarais-Gervais, 1994; Tremblay, Masse, Pagani, & Vitaro, 1996; Tremblay et al., 1992; Vitaro & Tremblay, 1994) combined parent training and child skill training. Primary school boys rated high on aggressive and disruptive behaviour (n = 166) were randomly assigned to a two year intervention or placebo control condition. Children worked with normative peers to develop more pro-social and adaptive social behaviour, while parents worked with family consultants approximately twice a month for two years to learn positive discipline techniques and how to support their child's positive behaviour. Initial results did not reveal clear group differences. At the three year follow-up when the boys were age 12, the treatment group was significantly less likely than control boys to engage in fighting, be classified as having serious adjustment difficulties, and to engage in aggression or delinquent activity. These results came from a variety of self, teacher, peer, and parent report measures. Effects of the treatment on other forms of antisocial behaviour (e.g., self-reported stealing) and substance use continued into early adolescence. Other early intervention programs have found durable effects which did not emerge until follow-up assessments (see Dadds et al., 1997). It should also be noted that intervention effects were reported by multiple informants across multiple domains of adjustment (i.e., behavioural, social, school/academic).

The First Steps Program (Walker, Kavanagh et al., 1998; Walker, Stiller, Severson, Feil, & Golly, 1998) also intervenes with both parents and children, the latter having been identified at kindergarten for exhibiting elevated levels of antisocial behaviour. Families with an at-risk child receive a 6 week home intervention and children participate in a classroom-based, skill-building and reinforcement program that lasts two months. The program has been evaluated with 42 subjects in two cohorts using a randomized controlled design. Positive treatment effects were found for both adaptive and academic behaviour at post-intervention and at follow-up into early primary school. A replication (Golly, Stiller, & Walker, 1998) with a new sample of 20 kindergarten students has produced similar results. Comparable positive results have also been found for a program that targets students aged six to 12 exhibiting aggressive and disruptive behaviour, their parents, and the classroom (Pepler, King, Craig, Byrd, & Bream, 1995; Pepler, King, & Byrd, 1991). In this program, the parent training is optional. It is important to note that in this study, significant differences between intervention and control children were only found on teacher ratings. Parents failed to see significant behaviour changes in the intervention children.

The CPPRG (Conduct Problems Prevention Research Group, 1992) implemented Fast Track, a school-wide program that integrates universal, selective, and indicated models of prevention into a comprehensive longitudinal model for the prevention of conduct disorders and associated adolescent problem behaviours. A randomised-controlled trial of 50 elementary schools in four U.S. urban and rural locations is still underway. 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 family (e.g., home visiting, parenting skills, case management), child (e.g., academic tutoring, social skills training), school, peer group, and community interventions. Targeted children were identified by multi-gate screening for externalising behaviour problems during kindergarten. The target group consisted of children from schools in neighbourhoods with high crime and poverty rates and who displayed the most extreme behaviour problems (top 10% of children as reported on externalising behaviour measures). At present, evaluations are available for the first three years (CPPRG, 1999a; CPPRG, 1999b). There have been significant reductions in special education referrals and aggression both at home and at school for the targeted children. The initial results provide evidence for improved social and academic development, including lower sociometric reports of peer aggression, and improved observers' ratings of the classroom atmosphere in the intervention sample. Evaluations will continue through middle school as Fast Track adopts an ecological-developmental model that assumes that, for high-risk groups, prevention of antisocial behaviour will be achieved by enhancing and linking protective factors within the child, family, school, and community.Top of page

Summary: It can be seen that recent community trials have been conducted that use randomised-controlled designs to evaluate multi-component programs based on comprehensive ecological and developmental models of externalising disorders. There are a number of characteristics that appear to be associated with successful EI for externalising problems in children. These include: 1) early identification and intervention beginning not later than preschool or early primary school years; 2) incorporation of family-based intervention as a core target for change; 3) adoption of a comprehensive model that emphasises a broad ecology (child, family, school, community); 4) adoption of a longitudinal/developmental approach to risk and protective factors and windows of opportunity for intervention; and 5) use of a comprehensive mix of selected (e.g., poor neighbourhoods), indicated (identification of aggressive children), and universal (e.g., classroom program) strategies.

Thus, successful prevention/early intervention and treatment in middle childhood has been achieved with regard to both internalising disorders and externalising disorders. Some of the studies cited have been able to demonstrate long-term improvements for children up to two years post-intervention. The long-term success of these interventions has clear implications for a concomitant reduction in community costs and family distress. None of the above studies took measures of substance use disorders at follow-up. However, given their focus on early risk factors for psychopathology, it is reasonable to speculate that these interventions have some potential for reducing the incidence of depression and substance use disorders in the adolescent years.

Adolescence, internalising disorders

Convincing literature points to the effectiveness of brief psychological interventions for internalising disorders in adolescents. However, the community impact of these brief programs is less convincing because of the low referral rates for internalising disorders during adolescence. The majority of adolescents in need of treatment simply do not receive it (Tuma, 1989). Thus, broader identification, recruitment, early intervention and prevention strategies become particularly important. Prevention of anxiety disorders in adolescence has received limited attention, although it should be noted that the treatment and prevention studies by Kendall, Barrett, Dadds, and Silverman reviewed above all included children up to 14 or 16 years in their successful reductions in anxiety disorders. Stress Inoculation Training Programs, which use a similar intervention to the anxiety treatments, have been shown to reduce anxious symptomology in universal adolescent samples (Kiselica, Baker, Thomas, & Reedy, 1994), as well as children evaluated to be at risk due to family breakdown (Pedro-Carroll, Alpert-Gillis, & Cowen, 1992).

In later adolescence, the pressing nature of such life threatening issues as depression, suicide, drug and alcohol abuse, or safe sex practices come to the forefront. With respect to internalising problems, the prevention of depression has gained more prominence than anxiety prevention in research investigations. This trend is in keeping with the proposals put forward in this chapter concerning a possible developmental pathway to substance use disorders from anxiety disorders in younger children, through depression as children move into adolescence, to substance use disorders. Thus, in this section the primary focus is on reviewing relevant programs for the prevention of depression.

To date, one of the most successful programs for reduction of depressive symptoms in young people has been the Pennsylvania Depression Program for adolescents aged 10 to 13 years (Jaycox, Reivich, Gillham, & Seligman, 1994). The study included three separate programs focusing on teaching (a) cognitive skills, (b) social problem solving skills, and (c) a combination of cognitive and social problem solving skills. Training in assertiveness, negotiation, and coping skills were also included. After finding no significant difference between the three intervention modalities, the groups were combined, resulting in a treatment sample of 69 participants and a wait list control group of 74 participants. Significant improvements in depressive symptoms were obtained for the intervention group compared to controls at post-testing, 6 month follow-up, and 2 year follow-up (Gillham, Reivich, Jaycox, & Seligman, 1995). This innovative study indicates that psycho-educational prevention efforts to build resilience to depression seem promising during early adolescence. A limitation of the study was the possible biasing effect of a self-selected sample in conjunction with the low initial recruitment rate (between 13% and 19%) and high attrition rate (30%).Top of page

In a second innovative study using an adaptation of the tertiary treatment approach developed by Lewinson et al. (1990), Clarke et al., (1995) reported significant improvements in depression for an indicated intervention group compared to a wait list group for 14 to 15 year old adolescents. The program was more successful than the Jaycox et al. (1994) study at recruiting adolescents. However, it still only succeeded in engaging less than 50% of the adolescents identified as being at risk for depression. There was also a reasonably high attrition rate, particularly in the intervention group (21 out of 76). In another indicated trial, Hains and Ellmann (1994) reported positive results for their program which consisted of problem solving, cognitive restructuring, and anxiety management, reducing depression scores in volunteer adolescents who had been classified as having high arousal levels. These authors also experienced difficulty with possible self-selection bias.

Beardslee and colleagues (Beardslee, 1989; Beardslee, Hoke, Wheelock, Rothberg, & et al., 1992; Beardslee & MacMillan, 1993; Beardslee, Salt, Porterfield, Rothberg, & et al., 1993) evaluated a selective program for adolescents and parents, where one or both parents had a major affective disorder, often in combination with other serious psychiatric disorders. The authors used family therapy and psycho-educational approaches to help families develop a shared perspective on the depressive illness, and to change parents' behaviour in relation to their children. In a controlled trial of 20 families, parents who received family-based interventions reported significantly more improvements in behaviour and attitudes than parents who received information alone. Recruitment was conducted through Medical Health Fund advertising, so no information is available regarding recruitment rates and self-selection processes.

The above studies provide evidence for the usefulness of selective and indicated prevention programs. They also highlight the well-known difficulties associated with recruitment and retention of adolescents. To the adolescent, such programs could be seen to single them out from the peer group at an age when peer group acceptance is especially important. This problem might be substantially reduced if intervention programs for adolescent depression could be implemented routinely as part of the school curriculum, as either an alternative or complement to indicated programs.

The Resourceful Adolescent Program (RAP: Schochet, Holland, & Whitefield, 1997) was developed to meet this need. It consists of components for adolescents (RAP-A) and their families (RAP-F). The RAP-A is a fully manualised 10 week group treatment (eight to 10 participants) focused on preventing depression through building adolescent resiliency. Given its universal delivery, participation rates approach 100% for the adolescents, although recruitment of families has remained a problem. Early results from controlled trials indicate that it is associated with reductions in self-reported depression, especially for adolescents with pre-existing depression at pre-treatment (Schochet et al., in press).

Summary: Thus, the evidence from adolescent groups is consistent with that from younger groups, supporting the efficacy of psychological skills building programs to reduce the incidence of internalising disorders in young people. It should be noted that the content of the anxiety prevention and depression prevention programs tends to be very similar, and includes core foci on cognitive skills, emotion regulation, dealing with challenges, and social problem solving skills. Unfortunately, none of the above studies has specifically measured substance use disorders as an outcome variable. Thus, the effect of these programs on reducing the prevalence of substance use disorders is at this stage unknown.

Adolescence, externalising disorders

The picture is somewhat different for externalising disorders, since adolescents who display various disorders consistent with externalising disorders represent a population at high risk for the development of substance use disorders, and therefore a population for whom intervention may well be beneficial. If such interventions are effective, reductions may not only be expected in recurrent prevalence of externalising disorders, but also in the incidence, prevalence and severity of substance use disorders (Bukstein, 1995). Of the preventive interventions for externalising disorders, particularly conduct problems, research indicates that behavioural family intervention has a high degree of efficacy both in the short term and after long-term follow-up (Prinz & Miller, 1994).Typically, behavioural family intervention will target parental interaction skills and parenting practice skills. In addition, a range of additional family risk factors will be addressed where warranted, for instance psychological state of parents (depression, anxiety, irritability), the presence of other identifiable marital problems, social support training, and the presence of substance use disorders.Top of page

Additional evidence for prevention of substance use disorders by intervening with internalising disorders and externalising disorders

The vast majority of substance misuse prevention studies for adolescents have focussed on externalising behaviours and social adversity risk factors. Several programs of research have now shown that reductions in externalising disorders can be effectively produced by the provision of skills building programs for the child, his or her family, and the school environment during the primary school years (see Greenberg et al., in press). Several of these studies have shown effective reductions in substance use disorders following targeting of externalising behaviour (e.g.,The Anger Coping Program: Lochman, 1992; Big Brother/Sister: Tierney et al., 1995).

There are a number of studies in which the promotion of general resilience in primary school children has been shown to reduce substance use into adolescence. For example, Schinke and Tepavac (1995) showed that a universal school-based intervention that focuses on personal and social decision making and assertive skills, reduced actual and potential substance use in eight to 11 year olds. The Seattle Social Development Project is a universal program that combines parent and teacher training throughout the primary school years. Controlled trials have compared early versus late scheduling of the intervention in large samples. Secondary school intervention was not effective. However, the early intervention model (i.e., targeting social competence in the primary school years and continuing across developmental phases) has been shown to effectively reduce substance use disorders at 18 years of age (Hawkins, Catalano, Kosterman, Abbott, & Hill, in press). Similarly, a number of well-designed studies that have targeted improved parent-child relationships have shown positive long-term benefits in terms of reductions or delays in drug taking (e.g., Kosterman, Hawkins, Spoth, Haggerty, & Zhu, 1997).

Several programs that aim to build skills and general resilience have been presented as selective programs. For example, Short (1998) reports on a preventive intervention for 10 to 13 year old children from divorced homes based on the rationale that coping skills mediate the effects of family stress on adolescent mental health and substance use. The intervention has been associated with improved coping and reductions in externalising, internalising, and substance use problems.

The overlap in skills focus between these programs and those aiming to reduce internalising disorders and externalising disorders is notable. That is, the focus on improving coping skills, problem solving skills and interpersonal relationships are common to most of the interventions. Also similar is their demonstrable positive outcomes, encouraging some optimism that the utilisation of school-based programs that increase resilience and reduce social and personal problems have the potential to reduce the development of substance use disorders.