V. Programs That Focus On Internalizing Behaviors

Diagnostic Criteria

Among the broad band of internalizing disorders, anxiety disorders and mood disorders are the most prevalent in childhood and have thus been the primary focus of preventive efforts. According to the DSM-IV (APA, 1994), mood disorders consist of depressive disorders (major depressive disorder (MDD), dysthymia (DD) and bipolar disorders (bipolar I, bipolar II). Diagnosis of a bipolar disorder in childhood is rare and MDD may begin at any age but the average age of onset is young adulthood (APA, 1994). It is estimated that prevalence rates for depression ranged from 0.4% to 2.5% for children and from 0.4% to 8.3% for adolescents (Birmaher, Ryan, Williamson, Brent, Kaufman, Dahl, et al., 1996).

Depressive disorders involve a pervasive mood disturbance in which the child or adolescent may experience sadness or irritability, a lack of interest or energy, hopelessness, feelings of worthlessness or inappropriate guilt, psychomotor agitation or retardation, or disturbance in sleep, appetite, or concentration (APA, 1994). In the past there has been controversy over whether young children are cognitively capable of depression. However, research has confirmed that depression is a clearly identifiable disorder in children (APA, 1994; Kovacs, 1996; 1997).

There are a number of diagnoses included within the Anxiety Disorders but the most typical diagnoses applied to children are Separation Anxiety Disorder, Social Phobia, Specific Phobia, and Generalized Anxiety Disorder (APA, 1994). Prevalence rates for these disorders range from 2.3 to 9.2% (Costello, 1989) and average around 8 percent (Bernstein & Borchardt, 1991). In general, anxiety disorders are characterized by excessive worry or distress. This may be in relation to a specific object or situation, or be more pervasive. Regardless of the source of their anxiety, for a child to be diagnosed with a disorder their distress must be significant enough to cause functional impairment. Additional symptoms include restlessness, poor concentration, irritability, sleep or eating problems, crying, or clinging (APA, 1994).


There is a high degree of comorbidity between and within the internalizing and externalizing dimensions. Harrington, Rutter, and Fombonne (1996) found that a significant number of depressed children and adolescents also develop a comorbid disorder. The most common conditions include dysthymia, anxiety disorders, disruptive disorders, and substance abuse. In the case of comorbid anxiety and depression in childhood, it appears that anxiety tends to precede depression (Kovacs, 1996). In addition to secondary diagnoses, other negative outcomes associated with depression and anxiety disorders include poor academic achievement, poor peer relations, and low self-esteem.

Current State of Knowledge

Although some recent research has focused on the developmental bases of childhood depression and anxiety, it is significantly less than that devoted to the disruptive behavior disorders (Birmaher et al., 1996; Cicchetti & Toth, 1998; NIMH, 1998; Spence, 1996; Spence & Dadds, 1996). As a result, developmental models of anxiety and mood disorders are only beginning to be posited and tested, and the diagnostic criteria for these disorders have failed to incorporate a developmental perspective. Risk and protective factors have been identified for both disorders based on etiological models and correlational studies. There are multiple pathways to each outcome and characteristics of the child are thought to interact with environmental and genetic influences in a complex manner. For example, Kagan and his colleagues (Kagan, Reznick, & Gibbons, 1989) have identified a stable temperamental pattern (i.e. behavioral inhibition) that is related to anxiety problems later in childhood (Biederman, et al., 1993). However, many behaviorally inhibited children do not develop anxiety disorders and researchers are only beginning to understand the mechanisms linking risk factors to the development of disorder.

Having a parent with either an anxiety disorder or depression increases the child’s risk of developing a similar disorder (Beardslee & Wheelock, 1994; Downey & Coyne, 1990; Last, Hersen, Kazdin, Francis, & Grubb, 1987; Mattison, 1992). In addition, sub-clinical levels of depressive symptomology are associated with increased risk of developing a depressive disorder (Gotlieb, Lewinsohn, & Seeley, in press) and children exhibiting early anxiety symptoms are considered at risk for developing adult anxiety disorders (Spence & Dadds, 1996). It is not clear how cognitive deficits and distortions play a role in the development of internalizing disorders or, why some children develop these maladaptive styles and other do not, but children who are depressed or anxious seem to have more difficulty in this area. Depressed children and adolescents exhibit impaired problem-solving abilities and a pessimistic or irrational cognitive style that impacts their perceptions (Beck, Rush, Shaw, & Emery, 1979; Quiggle, Garber, Panak, & Dodge, 1992). They also report feeling a lack of control over their lives and they are more likely to have depressive and hostile attributional biases (Kaslow, Brown, & Mee, 1994; Quiggle et al., 1992). Anxious children tend to have distorted perceptions of the degree of threat present in certain situations and lack the self-efficacy or effective coping skills to manage their internal distress. Thus, social-cognitions especially regarding self-efficacy, self-control, and cognitive distortion have become a focus for preventive interventions.

Negative life events such as the death of a parent, parental separation or divorce, or psychological trauma (e.g. exposure to violence, natural disaster) appear to play a causal role in the development of internalizing disorders (Goodyer & Altham, 1991). Certain family characteristics may place children at higher risk for these disorders either through genetic transmission or, as social learning theory would suggest, through social modeling and reinforcement. The work of Barrett, Rapee, Dadds, and Ryan (in press) suggests that compared to nonclinic parents, anxious parents may teach their children to perceive threat in ambiguous situations and utilize avoidant solutions to solve social problems. Khrone and Hock (1991) suggest that overcontrolling or overprotective parenting practices contribute to childhood anxiety by negatively impacting children’s ability to effectively learn problem-solving skills. Marital conflict and low cohesion have also been associated with elevated levels of depression in children (Fendrich, Warner, & Weissman, 1990).

Two protective factors that appear to modify the risk factors for internalizing disorders are social support and problem-focused coping strategies (Compas, 1987). Positive coping skills are associated with decreased levels of anxiety and distress and the cognitive difficulties associated with both disorders have the potential to undermine effective coping. Adaptive coping is also modeled through parent-child interactions. This may be another way that offspring of depressed or anxious adults are vulnerable for developing similar disorders as their parents.

Effective Preventive Interventions: Internalizing and Stress-Related Programs

Only one universal prevention program has targeted childhood depression and it failed to find significant effects (Clarke, Hawkins, Murphy, & Sheeber, 1993). Typically, prevention efforts in this domain are targeted towards the two groups considered most at-risk for developing the disorder: children of depressed adults and children or adolescents with elevated depressive symptomatology. Beardslee and his colleagues are currently conducting one of the most in-depth programs for children with depressed parents (Beardslee, Hoke, Wheelock, Rothberg, van de Velde, & Swatling, 1993; Beardslee, Salt, & Porterfield, 1992). Although their intervention model is promising, data regarding the effect of the intervention on children’s behavior is unavailable at this time.

Mood disorder prevention programs Prevention programs that target youth with elevated symptomatology are typically cognitive-behavioral and focus on the cognitive deficits and distortions associated with the disorder. The interventions typically take place in schools and are administered to students screened and selected from the general population. Clarke and his colleagues (Clarke, Hawkins, Murphy, Sheeber, Lewinsohn, & Seeley (1995) attempted to prevent unipolar depression in a sample of high school students with their 15-session, Coping with Stress Program. Subjects who endorsed scores of greater than 24 on the Center for Epidemiologic Studies – Depression Scale (CES-D; Redloff, 1977) were eligible to participate if they did not meet criteria for a depressive disorder. Although there were no significant differences at the end of the intervention, survival analyses that included assessments through 12 months post-intervention indicated that there were significantly fewer cases of MDD or Dysthymia in the experimental condition compared to the controls (p<.05).

The Penn Prevention Program is also directed toward altering the cognitive distortions and improving coping skills in at-risk youth. The participants in this program were younger than the Coping with Stress program, and considered at-risk for developing a depressive disorder due to elevated depressive symptoms (mean score on the CDI was 9.1 at pretest) and elevated levels of child-perceived family conflict. Students with elevated scores on both of these measures (greater than .50 summed z-scores) were eligible to participate in the program. Results from a quasi-experimental evaluation study suggested that the program resulted in clinically significant reductions in depressive symptoms immediately post-treatment and at a 6-month follow-up period. There was support that the reduction was mediated by changes in the children’s explanatory styles. Although there were no group differences in externalizing behavior post-treatment, at follow-up the parents of the intervention subjects reported significant improvements in children’s home behavior compared to controls. The intervention appeared to be most effective for subjects from high conflict families and those with high levels of depressive symptoms (above the median).

Anxiety prevention programs There have been few preventive efforts to reduce anxiety disorders in childhood and little research evaluating the effectiveness of those that have been attempted (Spence & Dadds, 1996). Programs directed towards reducing anxiety in medical procedures were not included in this review as they are less relevant to long-term risk models for the development of mental health disorders. Only one anxiety prevention program met our criteria.

The Queensland Early Intervention and Prevention of Anxiety Project is a large-scale, longitudinal study of a cognitive-behavioral, school-based program. It is designed to prevent the onset and development of anxiety problems in children by teaching them to utilize cognitive, behavioral, and physiological coping strategies while exposing themselves to increasingly fearful situations. The program is primarily focused on the individual child but includes three sessions with parents. The evaluation utilized a multi-level and multi-informant screening procedure to identify youth age 7 to 14 with elevated anxiety symptoms and youth who met the criteria for an anxiety disorder but in the less severe range. Schools were randomly assigned to experimental or control conditions. One of the difficulties in interpreting the findings of the evaluation are that the subjects with diagnoses and those with sub-diagnostic levels of anxiety were combined in many of the analyses. However, 6-months post-intervention when the anxious but non-disordered subjects who received the program were compared to controls they had developed significantly fewer internalizing disorders. This intervention effect was not apparent post-treatment.

Suicide prevention programs Other prevention programs have targeted internalizing symptoms that are related to depression and anxiety disorders. For example, depression is one of three behavioral risk factors (i.e. suicide-risk behaviors, depression, and anger) identified by Eggert and her colleagues as central in the prediction of suicidal potential in youth (Randell, Eggert, & Pike, in press). In a recent evaluation of two school-based intervention programs to prevent suicide, the authors demonstrated that at-risk students benefited in a variety of ways from a brief assessment and resource identification program (C-CARE) and a more intensive 12-session group life skills training group (CAST). High school students (9th-12th grade) were designated as "at suicide-risk" if they were a potential dropout from school and if their responses on a suicide risk screen met specific criteria designated by the authors. The results of a randomized clinical trial indicated that all three groups (C-CARE, C-CARE plus CAST, and a "care as usual" control) exhibited significant decreases in suicide-risk behaviors and anger problems. Both groups of intervention subjects received the C-CARE component and as a group they reported significantly lower levels of depression and higher levels of self-esteem compared to subjects in the control condition. The CAST component was found to be most effective in impacting the personal protective factors and family factors that the authors described as mediating the impact of the three primary risk factors for suicide potential. Subjects that received the CAST component evidenced significant improvements in self-control, problem-solving abilities, and perceived family support compared to controls.

Stress-related prevention programs Some intervention programs are not focused on preventing a specific disorder but are designed to reduce "stress" as it represents a more general internalizing symptom or as it is potentially triggered by a life event. Selected populations such as those experiencing a stressful life event are more susceptible to developing some form of psychopathology at these vulnerable points in development.

Two sets of researchers have applied the Stress Inoculation Training (SIT) counseling paradigm to prevention programs. Hains and his colleagues (Hains, 1992; Hains & Ellmann, 1994; Hains & Szyjakowski, 1990) designed a school-based prevention program (Stress Inoculation Training I) to reduce "negative emotional arousal" and other psychological problems associated with stress. They evaluated their program in a series of studies with high school students. In the most recent evaluation (Hains & Ellmann (1994), the sample consisted of youth exhibiting a combination of anxiety, depression, and poor anger control. After participating in a 13-session program that emphasized cognitive restructuring, problem solving, and anxiety management, students with the higher levels of stress before the intervention reported the most significant changes in anxiety and depressive symptomatology. These changes were considered clinically significant.

Kiselica, Baker, Thomas, & Reedy (1994) also applied their 8-session version of an SIT prevention program (Stress Inoculation Training II) to a sample of adolescents with elevated self-reported anxiety. Their intervention involved teaching the youth cognitive coping skills and assertiveness training. Intervention subjects reported significantly less anxiety and stress at post-test and 4 weeks after the intervention. While the design of the evaluation was good (i.e. randomized trial), the small sample size and restricted follow-up reduced the generalizability of the findings.

The Children of Divorce Intervention Project (Alpert-Gillis, Pedro-Carroll, & Cowen, 1989; Pedro-Carroll, Alpert-Gillis, & Cowen, 1992; Pedro-Carroll & Cowen, 1985) is an example of a school-based prevention program that is designed to prevent potential mental health complications that may result from parental divorce. CODIP is designed to provide children with a supportive outlet to discuss their parent’s divorce. It also focuses on counteracting any unrealistic perceptions or beliefs they may have about the divorce and build problem solving and adaptive coping skills. The program has been modified for different populations and evaluated on multiple occasions by the primary authors. The results indicate that based on responses from multiple informants (i.e. child, parent, teacher) the program is successful in improving children’s adjustment.

The Children of Divorce Parenting Program (Wolchik, West, Westover, Sandler, Martin, Lustig, Tein, & Fisher, 1993) is also a program that is designed to prevent the potential negative effects of divorce on children, but it focuses solely on parents. The program attempts to improve the quality of the custodial parent-child relationship by supporting parent’s effort to be warm and responsive, teaching parents to use clear and consistent discipline practices, and helping parent’s use anger management skills to reduce inter-parental conflict. In addition, parents are made aware of the importance of the father-child relationship and non-parental adults as a source of social support for the child. The program has only been evaluated on one occasion, and this was by the program developers. The results indicated that parents who participated in the program felt that their child evidenced significantly fewer problem behaviors at the end of the program compared to children in the control group. Forty-three percent of this main effect was mediated by improvements in the parent-child relationship as hypothesized by the authors.

The Family Bereavement Program (Sandler et al., 1992) is an intervention for children who have recently experienced the death of a parent. The program helps the entire family manage the grief process through education and social support. Children are taught adaptive coping skills and parents are encouraged to problem solve how to reduce stress, plan stable positive events, and maintain involvement with family members. The program was evaluated by its developers (Sandler et al., 1992) and the findings suggest that the intervention was successful in reducing children’s depressive symptoms and conduct problems according to parent’s report.

Programs with secondary effects on internalizing symptoms Although not intended primarily to reduce internalizing problems, two programs have shown significant reductions in this domain. The PATHS program (Kuche & Greenberg, 1994) was described in the Universal Programs section but is mentioned here as well. Evaluations of the PATHS curriculum conducted with elementary school aged children have shown significant decreases in depressive symptoms on self-report rating scales, and general internalizing symptoms by teacher report.

The Primary Mental Health Project (Cowen, Hightower, Pedro-Carroll, Work, Wyman, & Haffey, 1996) has also demonstrated significant impact on children’s internalizing symptoms. PMHP seeks to prevent psychopathology by providing additional targeted support to early elementary-age children who have been identified as having social/emotional or learning difficulties. The program uses a cadre of paraprofessional support staff coordinated by a school-based mental health professional in order to maximize the number of students who receive support. The intervention focuses on the school domain and changing both the school ecology and the individual child.

Based on an initial universal screening, children who are experiencing behavioral or learning difficulties are identified for PMHP services. The core intervention component is the development of an ongoing interactive relationship with a trained paraprofessional child associate. The child associate meets with the child alone or in small groups in a structured playroom equipped with items designed to encourage expressive play. The expression and exploration of all emotions is encouraged, with limits placed on inappropriate behavior. Child associates exploit opportunities for teaching life skills such as taking turns, following rules, and attending to a task.

In an early non-randomized study of PMHP with approximately 200 subjects, children who had successfully completed one year of the intervention were found to have significantly better adjustment on 2 separate teacher-rated measures of acting out, moody-withdrawn behaviors, and learning difficulties at post test than a matched control group or a group of students who had not successfully completed the intervention (Lorion, Caldwell & Cowen, 1976). With the exception of acting out behavior, these effects retained significance at 12 month followup.

Another non-randomized study of approximately 240 subjects found similar effects, with the greatest impact on students who began the program exhibiting more internalizing symptoms and less acting-out behaviors (Cowen, Gesten & Wilson, 1979). The findings were again based on teacher ratings. A third study without a control group measured pre- to post changes within seven consecutive annual cohorts ranging in size from 206 to 464. This study found significant improvement on 21 adjustment measures reported by teachers, paraprofessionals, and school-based mental health professionals. Though the PMHP has had few methodologically sound evaluations, the cumulative findings of over 20 years of evaluation have strongly supported the effectiveness of the program.

Promising programs In addition to the preventive interventions described above, this review identified two other programs that are promising. The Zuni Life Skills Curriculum (LaFromboise, 1991) was designed to reduce risk and build social-emotional competence in Zuni adolescents at risk for suicide and other self-destructive behavior. The intervention was evaluated using a quasi-experimental design (LaFromboise & Howard-Pittney, 1995) with students who participated in the intervention reporting significantly less hopelessness compared to non-intervention students. The intervention students were also observed as having higher levels of suicide intervention skills. While these findings are promising, the Zuni Life Skills Curriculum needs to be evaluated further as there were many aspects of the student’s functioning that were not affected by the intervention and liberal significance levels were used to test for group differences.

The Child Support Group (CSG; Stolberg & Mahler, 1994) is a 14-week, school-based program for children of divorce. The program focuses on building social-emotional and problem solving skills in the children. It also includes four parental workshops and materials for families to use at home. The intervention was evaluated on sample of 3rd through 5th grade students and included children of divorce and children from intact families. Children who met criteria for a DSM-III diagnosis were included in the sample. Schools were randomly assigned to the intervention or control conditions. While students who received the intervention were less likely to have adjustment problems (elevated TRF and CBCL Sum t-scores) at post-test and follow-up, effects were stronger for children with clinical-level adjustment problems at pre-test.