Keywords

Introduction

The primary purpose of this chapter is to present and critically evaluate current family-based treatments (i.e., parent management training [PMT] and other evidence-based approaches) for conduct problems (CP) in children and adolescents (collectively, we will refer to them as youth).Footnote 1 Family-based approaches to intervention have been applied to a wide variety of child problems and populations [e.g., attention-deficit/hyperactivity disorder (ADHD ; Abikoff et al., 2015); intellectual disability (Bagner & Eyberg, 2007); autism spectrum disorder (Bearss et al., 2015); anxiety (Cartwright-Hatton et al., 2011); depression (Eckshtain, Kuppens, & Weisz, 2017); child abuse (Vlahovicova, Melendez-Torres, Leijten, Knerr, & Gardner, 2017); and bullying (Healy & Sanders, 2014)], but it has the strongest and most extensive evidence base for children and adolescents with CP. Our focus is on the developmental period between ages 3 and 18 (i.e., preschool through high school age). CP can vary from annoying but relatively minor oppositional behaviors (e.g., yelling and temper tantrums) to more serious forms of aggressive behavior (e.g., fighting and physical destruction). In adolescence, youth may engage in certain types of CP that are illegal, and which are referred to as delinquent behaviors.

The first section of this chapter describes the theoretical underpinnings of CP and key family factors and processes in the development and maintenance of CP, as well as brief descriptions of selected family-based interventions for CP with children and adolescents. We then summarize the extensive evidence base for family-based interventions for child and adolescent CP, with discussion of both its strengths and limitations. The chapter concludes with suggestions for future research, policy, and practice.

Theoretical Background

In this section of the chapter, we will address the theoretical underpinnings of CP and family-based interventions for the treatment of CP.

Conduct Problems

Diagnostic Criteria, Epidemiology, and Developmental Pathways

The Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association [APA], 2013) specifies two different diagnostic categories pertaining to youth CP: oppositional defiant disorder (ODD) and conduct disorder (CD). ODD is defined as a persistent pattern of defiance and hostility against authority figures (e.g., parents and teachers). The DSM-5 distinguishes three separate—yet interrelated—affective and behavioral-based dimensions of ODD: (a) angry/irritable mood (e.g., temper tantrums); (b) argumentative/defiant behavior (e.g., refusing to comply with requests from authority figures); and (c) vindictiveness (e.g., showing spite; APA, 2013). While these ODD dimensions all share associations with later CP and disruptive behavior, there is also support for the idea that they differentially predict child outcomes. For example, some of the ODD symptoms pertaining to affective reactivity (i.e., temper outbursts, touchy or easily annoyed, anger and resentment) may be especially predictive of later risk for emotional disorders (e.g., Ezpeleta, Granero, de la Osa, Penelo, & Domenech, 2012; Herzhoff & Tackett, 2016; Stringaris & Goodman, 2009). By contrast, the vindictiveness (or hurtful) dimension of ODD seems to be largely predictive of aggressive CD symptoms (Stringaris & Goodman, 2009); however, vindictiveness may not manifest in young children until they are at least 4 years old (Ezpeleta et al., 2012).

CD is defined as a basic violation of other people’s rights or the norms followed by society. Common CD symptoms include destruction of property, starting fights with other youth, stealing, and truancy. A distinction is made between childhood and adolescent onset, with the presence of one or more CD symptoms prior to age 10 indicative of the former. Based on a burgeoning body of empirical evidence demonstrating the heterogeneity of CP (e.g., see Kimonis, Frick, & McMahon, 2014), DSM-5 has incorporated a specifier of “with limited prosocial emotions” to incorporate an approach to subtyping youth with CD according to the presence or absence of callous-unemotional (CU) traits (APA, 2013). CU traits are characterized by a lack of regard for other people’s feelings, deficient guilt associated with wrongdoing, restricted emotionality, and a lack of concern about poor performance at school, work, or in other significant activities. Youth with CD and clinically significant CU traits demonstrate more severe, chronic, and varied CP and antisocial behavior (Frick, Ray, Thornton, & Kahn, 2014).

Young children manifesting clinically severe levels of CP are more likely to meet criteria for ODD as opposed to CD. There is considerable evidence to suggest that ODD often precedes the development of CD in youth (e.g., Burke, Waldman, & Lahey, 2010; Rowe, Maughan, Pickles, Costello, & Angold, 2002); thus, many researchers consider ODD and CD to be age-related manifestations of a common syndrome (Lahey, Loeber, Quay, Frick, & Grimm, 1992), with CD representing a more severe developmental progression of CP (Loeber, Burke, & Pardini, 2009).

With respect to prevalence, the worldwide prevalence of ODD and CD among youth aged 6–18 years has been estimated to be 3.3% and 3.2% for ODD and CD, respectively (Canino, Polanczyk, Bauermeister, Rohde, & Frick, 2010). In general, boys are more likely than girls to display CP; however, this varies across different phases of development (Kimonis et al., 2014). For example, sex differences in ODD are minimal or nonexistent during preschool; however, during the school-age years, boys are 2–3 times more likely to be diagnosed with CP (ODD, CD) than girls. In adolescence, the rates increase for both boys and girls, and the sex gap diminishes somewhat.

In addition to early starters, longitudinal research sheds light on another distinct group of individuals who manifest high levels of CP in adolescence. The transition from childhood to adolescence is associated with increased engagement in risky and antisocial behaviors (e.g., drug use, stealing, truancy). Based on data from the Dunedin Longitudinal Study, Moffitt (1993) originally conceptualized a developmental pathway of CP that begins in adolescence but tapers off by early adulthood (i.e., adolescence-limited CP), and reflects exaggerated levels of somewhat normative adolescent behavior. However, follow-up analysis of these individuals showed persistence of antisocial behavior into their mid-20s and early 30s (Odgers et al., 2008). That is, their CP trajectory extended beyond adolescence and restricted their employment and educational opportunities, which in turn, likely contributed to other poor adult outcomes (e.g., substance use, physical health problems). In summary, CP can first manifest in childhood or adolescence; however, the early starter/life course persistent trajectory of CP is linked with a greater number of, and more severe risk factors, as well as more adverse outcomes across the lifespan (Fairchild, van Goozen, Calder, & Goodyer, 2013; Jolliffe, Farrington, Piquero, Loeber, & Hill, 2017).

In terms of comorbidity , many youth with either ODD or CD also manifest clinically significant symptoms of ADHD (Maughan, Rowe, Messer, Goodman, & Meltzer, 2004). The presence of ADHD is predictive of more negative outcomes (Kimonis et al., 2014). Youth with ODD and/or CD, especially if comorbid with ADHD, are at risk not only for the later development of more serious CP, but also for anxiety, mood and/or substance use disorders (e.g., Capaldi, 1991; Molina & Pelham Jr., 2003; Nock, Kazdin, Hiripi, & Kessler, 2007).

Conceptualizing Conduct Problems: Focus on the Family

The most comprehensive family-based formulation for the development of early-onset CP in children has been the coercion model developed by Patterson (e.g., Patterson, Reid, & Dishion, 1992). The model describes basic training in CP that occurs in the context of an escalating cycle of coercive parent–child interactions beginning prior to school entry. The proximal cause for entry into the coercive cycle is thought to be ineffective parental management strategies, particularly in regard to child compliance with parental directives during the preschool period. Types of parenting practices that have been closely associated with the development of child CP include inconsistent discipline, irritable explosive discipline, low supervision and involvement, and inflexible rigid discipline (Chamberlain, Reid, Ray, Capaldi, & Fisher, 1997). Recently, parental emotion socialization behaviors (such as emotion coaching, discussion of emotions, reactions to child emotions) have been implicated as small but significant predictors of concurrent and later child CP (Johnson, Hawes, Eisenberg, Kohlhoff, & Dudeney, 2017). Other family risk factors that may impact parenting practices include maladaptive social cognitions, personal (e.g., antisocial behavior, substance use, depression) and interparental (e.g., marital problems) distress, and social isolation (e.g., insularity; McMahon, Wells, & Kotler, 2006). Coercive interactions with siblings can also play a role in the development and maintenance of CP (Feinberg, Solmeyer, & McHale, 2012). Various child characteristics, such as comorbid disorders (e.g., ADHD, mood and anxiety disorders) and developmental phenomena (e.g., temperament, executive functions, emotion regulation, language development, social cognition) can also play a role in the development and maintenance of the coercive cycle (Greene, Ablon, Goring, Fazio, & Morse, 2004; McMahon et al., 2006).

Ineffective parenting and poor quality of parent–child relationship are also significantly implicated in the development and maintenance of adolescent CP. Among the various ineffective parenting practices associated with CP, poor parental monitoring is the strongest predictor of CP in adolescence (Racz & McMahon, 2011). In childhood, parental monitoring is largely restricted to the context of the home and school; however, in adolescence, youths’ increasing autonomy places more demands on parents to monitor the teenagers’ unsupervised activities with peers (especially those engaged in antisocial activities) and in the broader neighborhood. Seminal work by Stattin and Kerr (2000) showed that parents’ active efforts in monitoring their children, including their attempts to solicit information about, and control, their children’s activities, were less effective means of acquiring knowledge about their children’s whereabouts, compared with their children’s willingness to disclose this information. In other words, youth appear to be the gatekeepers of parents’ knowledge about them. Moreover, higher youth disclosure and greater parental knowledge are robustly associated with lower adolescent CP, whereas, paradoxically, parents’ increased attempts at soliciting information about their teens’ activities may be met with higher levels of CP over time (e.g., Kerr, Stattin, & Burk, 2010). Not surprisingly, adolescents with CP tend to disclose less than their peers without CP, thereby greatly restricting opportunities for parents to track, supervise, and set limits regarding their teens’ behavior and associations with deviant peers (Racz & McMahon, 2011). Importantly, youth may be more forthcoming about their behavior and peer associations when they share a warm and supportive relationship with their parents that facilitates open, spontaneous communication (e.g., Fletcher, Steinberg, & Williams-Wheeler, 2004). Taken together, these results suggest that when youth feel supported in the parent–child relationship, they show more willing disclosure, which, in turn, may increase parents’ knowledge and reduce risk for CP.

There is support for a cumulative risk conceptualization of serious CP and adolescent violence. For example, with respect to ODD, an increasing number of risks in the domains of parenting practices, child characteristics, attachment, and family adversity increase the likelihood of the development of ODD (e.g., Greenberg, Speltz, DeKlyen, & Jones, 2001; Harvey, Metcalfe, Herbert, & Fanton, 2011; Lavigne, Gouze, Hopkins, Bryant, & LeBailly, 2012). Dodge, Greenberg, Malone, and the Conduct Problems Prevention Research Group (CPPRG, 2008) substantiated a dynamic cascade model of risk factors, from early disadvantaged social context, to harsh/inconsistent parenting, to social and cognitive deficits, to CP behavior, to elementary school social and academic failure , to parental withdrawal of supervision, to deviant peer associations, to adolescent violence.

Family-Based Interventions for  Conduct Problems

Approaches to treating children with CP in the family have typically been based on a social learning-based parent management training (PMT) model of intervention (e.g., Miller & Prinz, 1990), whereas family-based interventions for adolescents have employed conceptually broader approaches (e.g., McCart & Sheidow, 2016).

PMT for Children with CP

The goal of PMT is to equip parents with behavior management techniques to improve the quality and consistency of their responding to both negative (e.g., defiance) and positive (e.g., compliance) child behavior. The envisaged outcome of PMT is a pattern of more positive parent–child interaction leading to an increased rate of child prosocial behavior and a reduction in CP. PMT is best practice for the treatment of CP in children (Kaminski & Claussen, 2017).

The underlying assumption of social learning-based PMT models is that some sort of parenting skills deficit has been at least partly responsible for the development and/or maintenance of CP. The core elements of the PMT approach include (a) intervention is conducted primarily with the parent or parent–child dyad, with relatively less therapist–child contact; (b) therapists refocus parents’ attention from a preoccupation with CP to an emphasis on prosocial goals; (c) the content of these programs typically includes instruction in the social learning principles underlying the parenting techniques; training in defining, monitoring, and tracking child behavior; training in positive reinforcement procedures including praise and other forms of positive parent attention and token or point systems; training in extinction and mild punishment procedures such as ignoring, response cost, and time-out in lieu of physical punishment; training in giving clear instructions or commands; and training in problem-solving; and (d) therapists make extensive use of didactic instruction, modeling, role playing, behavioral rehearsal, and structured homework exercises to promote effective parenting (Dumas, 1989; Kazdin, 1995; Miller & Prinz, 1990). PMT interventions have been successfully utilized in the clinic and home settings, have been implemented with individual families or with groups of families, and have involved some, or all, of the instructional techniques listed above. Furthermore, there is now substantial evidence that various forms of self-administered PMT (i.e., books, videos, internet-based interventions , smartphone apps) may be efficacious for some families (e.g., O’Brien & Daley, 2011; Watson MacDonell & Prinz, 2017) (Box 1).

Box 1 Is Time-Out an Appropriate and Effective Discipline Strategy?

As a significant component of PMT, time-out is associated with stronger treatment effects for CP (Kaminski, Valle, Filene, & Boyle, 2008). From the perspective of operant conditioning theory, time-out increases compliance (Owen, Slep, & Heyman, 2012) because it removes a child from reinforcers, including parental attention and fun activities, for a brief period of time when the child has misbehaved (Kazdin, 1980). Time-out is acceptable to parents participating in PMT (Cross Calvert & McMahon, 1987) and is used by the majority of parents in the community (Riley, Wagner, Tudor, Zuckerman, & Freeman, 2017; Tully et al., 1999).

Despite the large body of evidence for time-out, there has been increasing debate in the media about whether parents should be using time-out in response to child misbehavior (e.g., Siegel & Bryson, 2014). A common criticism of time-out is that it is ineffective for some children (Morawska & Sanders, 2011). Recent research regarding the real world implementation of time-out helps shed light on this issue. The proliferation of material about time-out over the Internet and on TV shows (e.g., Supernanny) has facilitated many opportunities for parents to learn about this discipline strategy. Findings from recent studies, however, highlight the negative influence of the media’s representation of time-out on parents’ behavior. For instance, although most parents (77%) report using time-out, the majority of them (85%) appear to be implementing it in ways that deviate from evidence-based practice (Drayton et al., 2017; Riley et al., 2017). This is not surprising considering that the vast majority of websites educating parents about time-out inaccurately describe its implementation, fail to include all of the research-supported components of time-out, or simply state that time-out is ineffective (Drayton et al., 2014).

Another common criticism of time-out is that it encourages an authoritarian style of parenting that may reject or psychologically isolate the child (Morawska & Sanders, 2011; Quetsch, Wallace, Herschell, & McNeil, 2015). Inherent in most evidence-based PMT programs is the fundamental approach of promoting a warm and positive parent–child relationship prior to implementing time-out contingent on child noncompliance and aggression. Time-out is only effective if time in—that is, time spent interacting with the parent—is more rewarding to the child. Advocates of time-out also argue that it serves as an emotion-regulation strategy for parent–child relationships in conflict situations (Webster-Stratton & Reid, 2017). Specifically, time-out may help scaffold children’s self-regulatory capacity and prevent parents from engaging in harsh discipline by interrupting the escalation of coercive parent–child interactions (Patterson et al., 1992). In this light, as a component of PMT, time-out may help prevent authoritarian parenting, including parental physical abuse (Chaffin, Funderburk, Bard, Valle, & Gurwitch, 2011).

In summary, claims that time-out is an ineffective and authoritarian response to child misbehavior are inconsistent with the conceptualization of appropriately implemented time-out and are not supported by empirical evidence. The widespread dissemination of time-out via social media may have both benefits and pitfalls; the majority of parents in Western cultures are now aware of this non-coercive form of discipline, but may lack understanding about its appropriate implementation.

We briefly describe several evidence-based PMT programs as examples of family-based treatments for children with CP. Descriptions of the clinical procedures utilized in these programs are widely available (e.g., therapist manuals, videotapes for therapist training, and/or books for parents), and each of the programs has been extensively evaluated.Footnote 2

The first three PMT programs have their origins in the pioneering work of Constance Hanf (see Kaehler, Jacobs, & Jones, 2016; Reitman & McMahon, 2013). They are (a) Helping the Noncompliant Child (HNC; McMahon & Forehand, 2003); (b) Parent-Child Interaction Therapy (PCIT; e.g., Zisser-Nathenson, Herschell, & Eyberg, 2017); and (c) The Incredible Years: BASIC Parenting Programs (BASIC; Webster-Stratton & Reid, 2017).Footnote 3 These Hanf-based PMT programs share common features. In general, they focus on treating noncompliance and other CP in younger children (i.e., preschool and early school age). Each of these interventions is divided into two phases. The primary goal of the initial phase is to break the cycle of coercive interactions by establishing a positive, mutually reinforcing parent–child relationship. In the second phase, parents are trained in giving clear and effective instructions to their children, and in implementing a systematic time-out procedure to decrease noncompliant behavior. HNC and PCIT are typically administered via individual contact with a therapist or trainer, whereas BASIC is designed primarily to work with parents in a group setting. Characteristic of all Hanf-based PMT programs, therapists make extensive use of modeling and role play during sessions (in addition to didactic instruction and discussion) to teach parents the skills of attends, rewards, ignoring, clear instructions, and time-out, and the use of home practice assignments and exercises. BASIC also employs a video/modeling group discussion format in which videos of parents interacting with their children in both appropriate and inappropriate ways are used as the impetus for discussion about appropriate ways to deal with child CP behavior. HNC and PCIT both use in vivo parent–child interactions for the purpose of coaching parents while they practice new parenting skills during session, which has been shown to augment the effectiveness of PMT (Kaminski et al., 2008). Similar to Hanf’s (1969) original program, two of the programs (HNC and PCIT) describe behavioral performance criteria that the parent must meet for each parenting skill.

The Triple P-Positive Parenting Program (Triple P; e.g., Sanders, 2012) has evolved over a 35-year period into a public health model for the promotion of healthy child and family functioning. Triple P comprises five levels of intervention, ranging from universal prevention strategies to an intensive and individualized treatment targeting children with severe CP symptoms. This model was designed for use with parents of children from birth to age 16, although the majority of outcome research has focused on families with young children (i.e., 2–8 years; Sanders, Kirby, Tellegen, & Day, 2014). Triple P interventions combine PMT strategies with a range of family support materials and services. Level 4 (Standard Triple P) is delivered in 8–10 sessions for parents of children with more severe CP symptoms. This level includes many components of traditional PMT programs such as a focus on parent–child interaction and training in parenting skills designed to be applicable to a range of problem behavior, and has been administered in individual, group, self-administered, and online formats. The Level 5 intervention (Enhanced Triple P) is appropriate when there is significant family dysfunction (e.g., parental depression, marital conflict) in addition to serious child CP. At this level, family-based intervention is individually tailored to families’ needs, and treatment strategies often include home visits focused on parenting practices, training in coping skills, and management of mood problems, marital conflict, and/or family stress.

The Generation Parent Management Training—Oregon (GenerationPMTO) program for preadolescent children (4–12 years of age) is described by Forgatch and Gewirtz (2017). Although most typically offered to individual families , GenerationPMTO can be delivered in a group format. In the individual format, children are incorporated into the sessions “as relevant” (Dishion, Forgatch, Chamberlain, & Pelham III, 2016, p. 820). Five core parenting skills are taught in GenerationPMTO: (a) skill encouragement (scaffolding using positive attention, incentive charts, and tangible rewards); (b) limit setting and discipline (e.g., time-out, response cost, fines, chores); (c) monitoring and supervision; (d) problem-solving (at the family level); and (e) positive involvement. The skills are taught sequentially, although the order may vary in the individual format. As in other PMT programs, significant emphasis is placed on in-session roleplaying and at-home practice assignments.

Family-Based Interventions for Adolescents with CP

In comparison to best-practice treatments for child CP that primarily focus on enhancing parents’ behavior management techniques (i.e., PMT), well-established interventions for adolescent CP target multiple risk factors in the family and other systems in which youth are embedded (McCart & Sheidow, 2016). This approach is based on a social-ecological model of the development of CP that posits interactional influences between youth and various family, peer, school, neighborhood, and community factors (Heilbrun, DeMatteo, & Goldstein, 2016). For instance, adolescents with serious and complex presentations of CP are more likely to have CU traits, a history of significant family disruption, gang affiliation, low school involvement, and involvement with juvenile justice (e.g., Frick et al., 2014; Kazdin, 1995; Kimonis et al., 2014). Although various environmental systems influence youths’ behavior, improving the quality of parent–child interaction continues to be a major goal in multimodal interventions for CP in adolescents. Below, we describe three different evidence-based psychosocial treatments for adolescent CP that have been evaluated in community settings, while focusing our discussion on the key family-based factors targeted by the programs. In the following section, we use the term family to refer to families headed by biological and foster parents.

Multisystemic Therapy (MST ; Henggeler, Schoenwald, Borduin, Rowland, & Cunningham, 2009) was developed as a treatment for adolescents aged 11–17 years with severe antisocial and delinquent behavior, and addresses risk factors (e.g., maladaptive parenting, deviant peer affiliation, poor school achievement) in multiple systems—including familial and extrafamilial—in which the adolescent is embedded. Intervention plans are tailored to individual cases, and designed in consultation with family members, based on a conceptualization of how risk and protective factors may be maintaining the adolescent’s CP. MST is delivered in the youth’s natural environment, such as during home and school visits, and leverages individual, family, and community resources to create support mechanisms that will maintain lasting behavioral change in the youth’s milieu. Parents are regarded as the linchpin of the intervention (Henggeler & Schaeffer, 2017), and the positive impact of MST on family relations is considered a key mechanism of change underlying improvements in youth CP (Huey, Henggeler, Brondino, & Pickrel, 2000). Treatment goals in the family domain include strengthening family structure and cohesion and parents’ behavior management practices. These objectives are achieved through implementing empirically proven strategies from various cognitive-behavioral (e.g., effective parental discipline and monitoring) and family (e.g., positive parent–teen communication, greater parental involvement in teens’ activities) therapies (Henggeler et al., 2009). Practitioners are available 24 h/7 days a week to provide immediate support for crises, and families typically receive 40–60 h of intervention over 3–5 months. Thus, MST is most cost-effective for youth referred by the juvenile justice system with serious CP.

Treatment Foster Care Oregon Model for Adolescents (TFCO-A ; Chamberlain, 2003) is a therapeutic model of foster care that provides intensive family- and community-based support for adolescents (12–17 years) with severe CP who may not receive appropriate services in typical foster care. As an alternative to group care, the therapeutic cornerstone of TFCO-A is the youth’s placement with specially trained foster parents who consult with members of a comprehensive treatment team (e.g., program supervisor, behavior support specialist, family therapist) regarding specific parenting strategies to manage the adolescent’s problematic behavior (Buchanan, Chamberlain, & Smith, 2017). These strategies are informed by social learning theory and include an individualized behavior modification system involving positive reinforcement and daily feedback, to promote youth adaptive behaviors (e.g., compliance with parents’ requests). The treatment team also provides individual therapy to adolescents, school- and community-based support, and crisis services, as needed. Adolescents’ biological parents are simultaneously involved in the intervention; they receive coaching in parenting strategies based on the PMT model (e.g., effective monitoring and consistent limit setting), which they begin to implement during home visits. Both foster and biological parents are considered significant agents of change in improving youths’ behavioral functioning (Buchanan et al., 2017). TFCO-A continues to support adolescents and their parents up to 3 months after family reunification, to prevent reentry into out-of-home-care.

Most empirically supported interventions for adolescent CP are based on cognitive-behavioral and family therapies (e.g., MST and TFCO-A). Considering that many adolescents with delinquent behavior have a history of adverse relational experiences (Stormo, Ortiz-Barreda, & Hollekim, 2017), and that attachment security can buffer risk for CP in adolescents with a maltreatment history (e.g., Joseph, O'Connor, Briskman, Maughan, & Scott, 2014), there is a need for trauma-sensitive interventions that largely focus on improving teens’ emotional bond with their parents. The Connect program (Moretti & Braber, 2013) was designed to strengthen attachment security in preteens and teens with serious CP, by shifting how parents understand, reflect on, and sensitively respond to the attachment meaning of their teens’ behavior (Moretti, Pasalich, & O’Donnell, 2015). Connect is delivered by two trained leaders who guide groups of 8–14 parents through ten 90-min sessions, each focused on an attachment principle that captures a key aspect of the parent–teen relationship and common parenting challenges (e.g., empathy, conflict, growth, and change). Experiential activities, including role plays and reflection exercises, are used to illustrate each principle and build parenting knowledge and skills. Specifically, the program enhances parents’ skills to promote secure attachment; sensitivity towards teens’ attachment needs (e.g., connection and independence); shared partnership with teens to strengthen collaborative problem-solving; and dyadic affect regulation to support teens’ management of difficult emotions. Although Connect may be suitable as a stand-alone intervention for adolescents with moderate levels of CP, it should be delivered in the context of a comprehensive treatment program targeting the various needs of adolescents with more severe cases of CP (Moretti & Braber, 2013).

Strengths and Limitations of the Evidence Base

The evidence base for family-based treatments is one of the largest and most impressive for any form of psychosocial intervention. Recent comprehensive reviews utilizing APA criteria for evidence-based treatments have identified PMT for children (Kaminski & Claussen, 2017) and certain family-based treatments for adolescents (McCart & Sheidow, 2016) as well-established (the highest level; i.e., MST and TFCO-A for the treatment of justice-involved youth) and probably efficacious (the second-highest level; e.g., MST for the treatment of CP in non-justice-involved youth).Footnote 4 A reflection of the extensive research base for family-based treatments can be seen in the large number of meta-analytic studies that not only address basic issues, such as treatment efficacy and effectiveness, but which have also assessed the evidence base for topics such as effective components (Kaminski et al., 2008; Lipsey, 2009), maintenance of treatment effects (van Aar, Leijten, Orobio de Castro, & Overbeek, 2017), implementation (Leijten, Melendez-Torres, Knerr, & Gardner, 2016; Michelson, Davenport, Dretzke, Barlow, & Day, 2013), moderators (e.g., Lundahl, Risser, & Lovejoy, 2006), and specific programs [e.g., Incredible Years (Menting, Orobio de Castro, & Matthys, 2013), PCIT (Ward, Theule, & Cheung, 2016), Triple P (Sanders et al., 2014); MST (van der Stouwe, Asscher, Stams, Deković, & van der Laan, 2014)]. Where possible, the following discussion of the status of the evidence base for family-based treatments will focus on findings from these meta-analytic studies.

Generalization and Social Validity

The short-term efficacy of PMT in producing changes in both parent and child behaviors has been demonstrated repeatedly (e.g., Comer, Chow, Chan, Cooper-Vince, & Wilson, 2013; Piquero et al., 2016; Sanders et al., 2014; Serketich & Dumas, 1996), but generalization of these effects is also important to demonstrate. Forehand and Atkeson (1977) described four types of generalization of PMT intervention effects: setting, temporal, sibling, and behavioral. There have been a number of investigations assessing the various types of generalization that have, for the most part, supported the efficacy of behavioral PMT programs.

Each of the PMT programs described earlier in the chapter has documented setting generalization from the clinic to the home for parent and child behavior and for parents’ perception of child adjustment (e.g., Fleischman, 1981; Peed, Roberts, & Forehand, 1977; Sanders, Markie-Dadds, Tully, & Bor, 2000; Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998; Webster-Stratton, 1984). Recent meta-analyses (Sawyer, Borduin, & Dopp, 2015; van Aar et al., 2017) have documented the temporal generalization of intervention effects for both PMT and other family-based interventions for at least 1 year post-treatment. In their meta-analytic review of PMT, van Aar and colleagues noted evidence for occasional sleeper and fade-out effects (i.e., increased improvement or deterioration following treatment, respectively). Individual studies conducted 4.5–14 years after completion of the HNC program suggest that the youth were functioning well compared to peers in a community comparison group (selected at the time of follow-up) in terms of parent-, teacher-, and self-reported adjustment (Forehand & Long, 1988; Long, Forehand, Wierson, & Morgan, 1994). Similar findings have been reported for the BASIC program at follow-ups ranging from 7 to 12 years (Scott, Briskman, & O’Connor, 2014; Webster-Stratton, Rinaldi, & Reid, 2011). Long-term follow-ups of Triple P at 10 years (aged 3–13 years; Hahlweg & Schulz, 2018) and 15 years (Smith, 2015) have also recently been reported. Positive long-term outcomes on reducing serious criminal outcomes have been reported for MST compared to individual therapy up to 21.9 years after treatment initiation (Sawyer & Borduin, 2011).

Several investigators have now assessed setting generalization from the clinic or home setting to the school. In their meta-analytic study, Serketich and Dumas (1996) reported an effect size of 0.73 for PMT when the outcome was based on teacher report, and McNeil, Eyberg, Eisenstadt, Newcomb, and Funderburk (1991) demonstrated generalization of PCIT to the classroom using both observational data and teacher ratings of CP behavior. However, other investigators have failed to find evidence of generalization to school or a failure to maintain this generalization (e.g., Breiner & Forehand, 1981; Taylor, Schmidt, Pepler, & Hodgins, 1998).Footnote 5

Several PMT programs (HNC, PCIT, GenerationPMTO, BASIC) have demonstrated sibling generalization (e.g., Brestan, Eyberg, Boggs, & Algina, 1997; Gardner, Burton, & Klimes, 2006; Horne & Van Dyke, 1983; Humphreys, Forehand, McMahon, & Roberts, 1978), and this generalization has been maintained up to a 1 year follow-up for GenerationPMTO (Horne & Van Dyke, 1983). Behavioral generalization from the treatment of child noncompliance to other behaviors (e.g., aggression, temper tantrums) has been demonstrated for HNC (Wells, Forehand, & Griest, 1980), BASIC (Webster-Stratton, 1984), and GenerationPMTO (e.g., Fleischman, 1981). Similarly, family-based treatment effects on comorbid disorders (e.g., ADHD, depression, anxiety) may be considered to be a type of behavioral generalization. For example, children who displayed comorbid ADHD/ODD and who participated in HNC improved in both domains (Forehand et al., 2016). In a recent review, Gonzalez and Jones (2016) reported on the cascading effects of PMT for comorbid child internalizing problems. Meta-analytic results from randomized controlled trials (RCTs) comparing MST against usual community care suggest that MST has small but significant effects not only on reducing adolescent CP but on comorbid psychopathology and substance use (van der Stouwe et al., 2014).

The social validity of PMT interventions with children with CP has been assessed by various methods, including measures of consumer satisfaction completed by parents (e.g., McMahon & Forehand, 1983), treatment acceptability (e.g., Cross Calvert & McMahon, 1987), and by determining the clinical significance of improvements (e.g., Sheldrick, Kendall, & Heimberg, 2001). PMT programs have provided strong evidence of consumer satisfaction at post-treatment and/or follow-up periods of a year or more (e.g., Brestan, Jacobs, Rayfield, & Eyberg, 1999; Leung, Sanders, Leung, Mak, & Lau, 2003; McMahon, Tiedemann, Forehand, & Griest, 1984; Patterson, Chamberlain, & Reid, 1982; Taylor et al., 1998). They have also provided normative comparisons indicating that, by the end of treatment, child and/or parent behavior more closely resembles that in non-referred families (e.g., Forehand, Wells, & Griest, 1980; Sanders & Christensen, 1985; Sheldrick et al., 2001). In their meta-analytic review of PMT, Serketich and Dumas (1996) reported that 17 of 19 intervention groups dropped below the clinical range after treatment on at least one measure, and 14 groups did so on all measures. Similarly, in a qualitative review of PCIT, Gallagher (2003) found clinically significant improvements (i.e., drop below clinical cutoff) in 14 of 17 studies.

There is also research to suggest that PMT can be acceptable and effective in culturally diverse families (e.g., Reid, Webster-Stratton, & Beauchaine, 2001). However, the extent to which interventions need to be systematically modified to be culturally relevant is unclear (Baumann et al., 2015; Gardner, Montgomery, & Knerr, 2016; Mejia, Leijten, Lachman, & Parra-Cardona, 2017).

It is apparent that evidence for the generalization and social validity of family-based interventions with children with CP is extensive and, for the most part, positive. Furthermore, such interventions have also resulted in positive changes in parenting stress and increases in perceived parenting competence following treatment (see Colalillo & Johnston, 2016, for a review). However, systematic changes in parental adjustment that were more distal from parenting (e.g., parental depression, marital functioning) were less clear.

Comparison Studies

Each of the family-based programs described above (and many others) have been positively evaluated compared with no treatment, waiting-list, or attention-placebo control conditions (e.g., Lundahl et al., 2006; Medlow, Klineberg, Jarrett, & Steinbeck, 2016; Piquero et al., 2016; Serketich & Dumas, 1996; van der Stouwe et al., 2014). Furthermore, comparisons with groups of non-referred typically developing samples have indicated greater similarity in parent/child behaviors and/or parental perceptions of children after PMT (e.g., Forehand et al., 1980; Patterson, 1974).

As evidence for the efficacy of various interventions with children with CP has accumulated, increased attention has been focused on the relative efficacy of these interventions compared to other forms of treatment. Several family-based treatment programs have been shown to be more efficacious than family systems therapies (e.g., Patterson & Chamberlain, 1988; Wells & Egan, 1988), the STEP program (Baum, Reyna McGlone, & Ollendick, 1986), couples coping enhancement training (Bodenmann, Cina, Ledermann, & Sanders, 2008), and available community mental health services (e.g., Patterson et al., 1982; Stattin, Enebrink, Ozdemir, & Giannotta, 2015; Taylor et al., 1998; van der Stouwe et al., 2014; Westermark, Hansson, & Olsson, 2010). Compared with group care , TFCO-A significantly reduced delinquency and deviant peer affiliations for boys and girls, and improved parenting outcomes and placement stability for boys (Dishion et al., 2016). Similar findings were demonstrated in a Swedish RCT of TFCO-A versus treatment as usual (Bergström & Höjman, 2015; Westermark et al., 2010).

Meta-analytic studies have demonstrated that PMT has stronger effect sizes than home visiting interventions (ES = 0.39 and 0.28, respectively) with young children (5 years old and younger; Piquero et al., 2016), and youth cognitive behavior therapy in decreasing CP (ES = 0.45 and 0.23, respectively) with 6- to 12-year-olds (McCart, Priester, Davies, & Azen, 2006). Recently, some researchers have reported comparable effects of other family-based interventions to PMT. For example, Duncombe et al. (2016) reported equivalent effects for the Tuning in to Kids program (which is an emotion-focused parenting program; Havighurst & Harley, 2007) to an 8-session version of Group Triple P with elementary school-aged children. Similarly, Ollendick et al. (2016) found comparable effects for Barkley’s (1997) Defiant Children (a Hanf-based program) and Collaborative and Proactive Solutions (Greene, 1998), which employs a problem-solving model with parents to address child ODD. Head-to-head empirical comparisons of different PMT programs have been conducted (e.g., Abikoff et al., 2015; Högström et al., 2017; Stattin et al., 2015). Two meta-analytic studies comparing PMT programs reported that, while all of the PMTs had positive effects, the effect sizes were larger for PCIT on some outcomes (e.g., child behavior change) than Triple P (Piquero et al., 2016; Thomas & Zimmer-Gembeck, 2007) and for BASIC (Piquero et al., 2016). Thomas and Zimmer-Gembeck suggested that providing opportunities for parent–child interaction within the session may have accounted for this difference, consistent with the findings of Kaminski et al. (2008) in their meta-analysis of PMT. In an RCT (N = 908 Swedish families) comparing Connect against three established PMT programs (including BASIC), Connect had treatment effects of a similar magnitude as the PMT programs at 2-year follow-up (Högström et al., 2017). However, CP outcomes immediately post-treatment slightly favored the PMT programs over Connect (Stattin et al., 2015).

Mechanisms and Moderation

Given that a core premise of PMT (and some other family-based treatments for adolescents such as MST and MTFC) is that change in parenting behavior is the active mechanism for producing child behavior change , it is surprising that this issue has only been addressed empirically fairly recently (Fagan & Benedini, 2016; Forehand, Lafko, Parent, & Burt, 2014). Forehand and colleagues identified 25 studies (all of them conducted since 2000) that examined one or more parenting behaviors as potential mediators of child and adolescent outcomes in family-based treatments. Less than half (45%) of the analyses supported mediation. This was most likely to occur for composite measures of parenting (90% supported mediation), discipline (55%), and positive parenting (45%), and least common for negative parenting (26%) and monitoring (10%). Mediation was more common in prevention as opposed to treatment studies (72% vs. 32%) and in samples of younger children (i.e., less than 10 years old; 61% vs. 29% for older children). Reasons for these findings are not known, but the authors speculate that mediation may be more likely with younger children whose behaviors are less entrenched, making the child’s behavior more amenable to parental influences. Other potential mediators have been examined even less frequently. Parenting sense of competence has been shown to mediate the effects of MST (Dekovic, Asscher, Manders, Prins, & Van der Laan, 2012). Reducing engagement with deviant peers is one candidate that has received support in both MST (Huey et al., 2000) and TFCO-A (Eddy & Chamberlain, 2000). Connect appears to decrease youth CP by way of reducing attachment avoidance and enhancing affect regulation in teens (Moretti, Obsuth, Craig, & Bartolo, 2015).

In general, there has been a relative dearth of attention paid to the extent to which family-based treatments may be differentially efficacious with different subgroups of children, parents, and families, or as a function of different aspects of PMT (e.g., treatment delivery mode). An early meta-analytic study that examined moderators of PMT found that more severe child CP, single-parent status, economic disadvantage (i.e., low socioeconomic status), and group-administered (as opposed to individually administered) PMT resulted in poorer child behavior outcomes (Lundahl et al., 2006). In addition, economic disadvantage and PMT alone (as opposed to multicomponent interventions that included PMT) were also associated with poorer parent behavior and parental perception outcomes. Child age was not a significant moderator, which has also been reported by others (e.g., McCart et al., 2006). Lundahl and colleagues found that among disadvantaged families, individual PMT was associated with more positive child and parent behavioral outcomes than group PMT. A qualitative review of 19 studies by Shelleby and Shaw (2014) concluded that the effects of PMT were quite robust across a variety of sociodemographic and family risk factors; however, in contrast to Lundahl et al.’s findings, higher levels of baseline child CP were associated with more positive outcomes from PMT. Family-based treatments appear to be comparably effective for boys and girls (Kaminski & Claussen, 2017; Leve, Chamberlain, & Kim, 2015).

Meta-analytic studies have examined potential moderators for Triple P, BASIC, PCIT, and MST. In a comprehensive meta-analysis of 101 studies focused specifically on moderators of Triple P, greater severity of child behavior problems (for the parental relationship outcome variable), study approach (targeted and treatment approaches had stronger effects on child behavior outcomes than universal approaches), and Triple P level (i.e., Triple P Levels 3, 4 and 5 vs. Level 1 moderated effects on treatment satisfaction and efficacy) were factors associated with larger treatment effects when controlling for other significant moderators (Sanders et al., 2014). A meta-analysis of 50 studies of BASIC found that initial severity of child CP was the most powerful moderator of post-treatment effects, with more severe CP behavior associated with more positive outcomes (Menting et al., 2013). Parental attendance at more sessions and receipt of BASIC alone (without other treatment components of the Incredible Years intervention package) were also associated with larger effect sizes. However, it is important to note that a recent trial of BASIC in the Netherlands, which employed a large sample (N = 387), both parent-report and observational outcome measures, and multivariate analyses, found minimal evidence of moderation, with only 3 of 40 tested moderation effects being significant (one of which was parental attendance; Weeland et al., 2017). A small meta-analysis (12 studies) of PCIT reported no moderation of intervention effect by child sex or diagnosis (ODD, CD, ADHD; Ward et al., 2016). Meta-analytic analyses suggest that larger MST effects have been obtained for adolescents younger than 15 years, Caucasian youth, and in US samples (van der Stouwe et al., 2014). The latter finding may be linked with challenges in implementing MST in countries outside of the US (e.g., poor treatment adherence), and to lower base rates and severity of offending behavior and higher quality usual care services than in the US (Asscher, Dekovic, Manders, van der Laan, & Prins, 2013; Henggeler & Schaeffer, 2017).

One area of current research interest is the extent to which family-based treatments are efficacious with a subgroup of children and youth with CP who also display CU traits. Children with CP and elevated levels of CU traits do not respond as well to traditional PMT interventions as do other children with CP. In a recent review, CU traits were associated with poorer outcomes from family-based treatments in 81% (9 of 11) of the studies (Hawes, Price, & Dadds, 2014). However, it is also the case that these children do respond to family-based intervention, but to a lesser degree than other children. Interestingly, this appears to be more likely to occur with children with an ODD diagnosis than with a diagnosis of CD (Hawes et al., 2014). Furthermore, four studies have documented decreases in CU traits (in addition to decreases in CP) as a function of family-based interventions (Butler, Baruch, Hickey, & Fonagy, 2011; Kjøbli, Zachrisson, & Bjørnebekk, 2018; McDonald, Dodson, Rosenfield, & Jouriles, 2011; Somech & Elizur, 2012). It has been suggested that additional emphasis be placed on the promotion of parental warmth and positive reinforcement in family-based interventions with these children (Hawes et al., 2014). Supporting such a recommendation are recent findings that changes in positive (but not negative) parenting mediated the effects of intervention on CU traits (Kjøbli et al., 2018; Pasalich, Witkiewitz, McMahon, Pinderhughes, & CPPRG, 2016).

Implementation

Large-scale effectiveness trials of PMT and other family-based treatments as well as cross-cultural dissemination studies have become common. These research efforts provide essential information on the feasibility of transporting interventions for CP to real-world settings and utilizing such interventions with diverse populations of children and families across the globe.

With respect to effectiveness, a meta-analysis demonstrated that PMT was more effective than waitlist control conditions when conducted in real-world settings, as indicated by: (a) clinic-referred samples; (b) non-specialist therapists; (c) routine settings; and (d) as part of a routine service (Michelson et al., 2013). Well-established family-based programs have been implemented in local community mental health centers (e.g., Henggeler, Melton, Brondino, Scherer, & Hanley, 1997; Scott, Spender, Doolan, Jacobs, & Aspland, 2001; Stattin et al., 2015; Taylor et al., 1998), volunteer organizations (Gardner et al., 2006), and in the child welfare/protection system (e.g., Chaffin et al., 2011; Chamberlain et al., 2008; Letarte, Normandeau, & Allard, 2010; Marcynyszyn, Maher, & Corwin, 2011).

Furthermore, many of these interventions have now been evaluated in international settings. Two recent meta-analytic reviews have demonstrated the transportability of PMT programs from their country of origin to other countries, both Western and otherwise (Gardner et al., 2016; Leijten et al., 2016). Gardner and colleagues reported effects of PMT in the destination countries comparable to those obtained in the program’s country of origin. Interestingly, effects were somewhat stronger in regions that were culturally more distant (e.g., Asia, Latin America, Middle East) as opposed to countries with Anglo/European roots (e.g., Canada, the UK, Ireland, Norway, Sweden). Leijten and colleagues compared the effectiveness of transported and homegrown PMT programs in four geographic regions (North America, Australia, English-speaking European countries, and other European countries). They found comparable effectiveness between homegrown and transported programs, regardless of the geographical region or the particular brand of PMT program (i.e., BASIC, PCIT, Triple P, GenerationPMTO). The authors suggest that these findings support both the dissemination of PMT programs to different countries, and the utility of locally developed programs that are based on similar principles (e.g., social learning) and that have been carefully evaluated. A potential limit to the generalization of these findings is that the regions included in these studies were, for the most part, high-income countries. Efforts to establish and evaluate PMT in low- and middle-income countries are just beginning (e.g., Knerr, Gardner, & Cluver, 2013; Mejia, Calam, & Sanders, 2012).

Evaluations of family-based treatments for adolescents have also been conducted in international settings, including MST [Canada (Cunningham, 2002), the Netherlands (Asscher et al., 2013), Norway (Ogden & Amlund-Hagen, 2006), Sweden (Sundell et al., 2008)], TFCO-A [Sweden (Bergström & Höjman, 2015; Westermark et al., 2010); the UK (Sinclair et al., 2016)], and Connect (Sweden; Högström et al., 2017; Stattin et al., 2015). Whereas findings for TFCO and Connect have generally been positive, this is less so for MST. As noted above, this may be at least partially due to less severe offending patterns and higher levels of usual treatment services for offending adolescents in the destination countries (Henggeler & Schaeffer, 2017).

Economic Analyses

It is well-established that children with CP, especially those who follow the early-starter developmental pathway, have the potential to incur substantial societal and economic consequences. For example, it has been estimated that the potential value of saving a single high-risk youth from a criminal career ranges from US$3.2 to $5.5 million (Cohen & Piquero, 2009). Given these figures, PMT and other family-based interventions have great potential to provide a cost-effective means of preventing future delinquency and perhaps even adult criminal activity. To date, there have been relatively few empirical examinations of cost-effectiveness (for reviews, see Charles, Bywater, & Edwards, 2011; Christenson, Crane, Malloy, & Parker, 2016). Some of the most thorough and methodologically sophisticated analyses have been conducted by the Washington State Institute for Public Policy (WSIPP, 2017). These analyses suggest benefit-to-cost ratios ranging from US$1.79 to US$3.36 for IY, PCIT, HNC, GenerationPMTO, and Triple P, and US$2.42 for MST and US$2.08 for TFCO-A (dollar values greater than 1 indicate that the benefits of a program exceed its costs). In addition, cost savings may be even greater when coordinated, multilevel systems of intervention are implemented. For example, WSIPP estimated that implementation of the Triple P system at a population level was associated with a benefit-to-cost ratio of US$9.17.

Future Directions for Research

It is apparent that the evidence base for family-based interventions for the treatment of youth CP is extensive and growing. Future research should continue to focus on extending this research base in the areas covered in the previous section of this chapter (i.e., generalization and social validity, comparisons with other treatments, mechanisms and moderation, implementation in real-world settings with diverse populations of children and families, and economic analyses).

With respect to mediation, as noted above, the research base has been primarily limited to a relatively small number of studies that have examined parenting practices as potential mediators. Future research should include parallel testing of multiple mediators (Patel, Fairchild, & Prinz, 2017) and more complicated mediational pathways, for instance, involving sequential or cascading effects (e.g., Forehand et al., 2014; Sandler, Schoenfelder, Wolchik, & MacKinnon, 2011). Analyses of moderated mediation and mediated moderation can also be employed to modify existing interventions or to develop new ones (Fagan & Benedini, 2016). Moreover, these more complex models have potential for informing developmental theory on the interplay of risk and protective factors, by examining whether a developmental cascade of risk factors associated with poor child outcomes (mediation pathway) may be mitigated by assignment to a family-based intervention versus control (moderator; e.g., Pasalich, Fleming, Oxford, Zheng, & Spieker, 2016).

Several important areas for future research on family-based interventions for children with CP can be subsumed under the label of personalized mental health interventions (Ng & Weisz, 2016, 2017), which are “evidence-based methods for matching and tailoring treatments to individuals to optimize their outcome” (Ng & Weisz, 2017, p., 503). One approach is to modify treatments based on particular characteristics of children (e.g., CU traits, comorbid anxiety) and/or families (e.g., foster families, military families). Initial explorations of the roles of neuroendocrine functioning (e.g., Shenk et al., 2012) and gene by treatment interactions (e.g., Chhangur et al., 2017) in predicting or moderating treatment outcome represent exciting avenues for potentially improving family-based treatments for youth with CP. For example, Chhangur and colleagues recently documented that boys (but not girls) carrying high numbers of dopaminergic plasticity genes demonstrated greater decreases in parent-reported CP behavior as a function of parental participation in the BASIC PMT program.

Another approach to personalizing intervention that has received increased attention is the embedding of common elements of evidence-based interventions into modular treatment protocols (e.g., MATCH; Weisz et al., 2012). In essence, therapists select various intervention components that have empirical support in the treatment of different child disorders (e.g., time-out, response prevention, exposure to anxiety-eliciting stimuli), rather than relying on a set package of intervention techniques from a named program for a single child disorder. This approach has particular promise for therapists working with clinic-referred children, who typically present with multiple disorders, and enhances therapist flexibility in terms of offering a menu of evidence-based components and a sequence of decision rules for implementing them. On another front, common elements for PMT programs have been identified as well (Barth & Liggett-Creel, 2014; Kaehler et al., 2016). Recent findings suggest that modular treatment for youth mental health may be more effective than community-implementation of evidence-based treatments (Chorpita et al., 2017).

A third approach to personalization is a focus on the processes of parental engagement with family-based interventions, which typically includes attendance, adherence (e.g., in-session participation, homework completion), and cognitions (e.g., agreement with treatment rationale, therapeutic alliance, treatment satisfaction; for reviews, see Chacko et al., 2016; Nock & Ferriter, 2005; Piotrowska et al., 2017). A recent review of 262 PMT studies by Chacko and colleagues found a combined attrition rate of 51% (failure to enroll in or to complete treatment). Lower socioeconomic status was associated with higher attrition. There was a paucity of data concerning the other elements of engagement. The authors note the need for uniformity in reporting the different forms of engagement, including strategies designed to facilitate engagement. While there has been increasing attention to developing and evaluating such strategies (e.g., Chacko et al., 2016; Ingoldsby, 2010; Nock & Kazdin, 2005), additional research in this area is sorely needed. The recent presentation of a comprehensive process model of engagement (CAPE; Piotrowska et al., 2017) provides an excellent heuristic framework for future research in this area. The elements include Connect and Attend (i.e., enrolment and attendance), Participate (which includes in-session discussion and homework completion), and Enact (implementation of the newly learned parenting strategies). Relatedly, others have called for the need for research focused on skill acquisition and utilization in the treatment of youth CP (Lindhiem, Higa, Trentacosta, Herschell, & Kolko, 2014).

Personalizing intervention can also relate to how family-based treatments are delivered. Prior research described in this chapter has indicated some of the relative advantages and disadvantages of individual versus group administration of family-based interventions and the value of self-administered treatments (using a variety of formats) for certain families. For example, group-based PMT can be a cost-effective alternative to individual family treatment in some instances, and may ultimately have a greater impact at the community level, given the ability to reach larger numbers of families. However, PMT conducted with individual families may be more efficacious with economically disadvantaged families (Lundahl et al., 2006). In addition, there is some evidence that child participation in PMT sessions is associated with more positive outcomes (Kaminski et al., 2008; Kaminski & Claussen, 2017). A recent review concluded that brief PMT interventions (i.e., eight or fewer sessions) may be sufficient for reducing child CP in some families (Tully & Hunt, 2016), and Bagner and colleagues (Bagner et al., 2016; Bagner, Garcia, & Hill, 2016) have shown that an adapted version of PCIT [primarily the initial phase of treatment (Child-Directed Interaction)] can enhance parent–child relationships, reduce CP, and improve language production in 12- to 15-month-old infants. It is worth noting that one advantage of the Triple P multilevel system of intervention is that it allows for customization of program and titration of dose based on problem severity, mode of delivery, and parental preference.

Space limitations preclude a thorough discussion of the burgeoning research on the development and evaluation of technology-based interventions, which include both stand-alone and technology-enhanced interventions. The former refers to those technology-based interventions that do not involve any clinician contact (e.g., self-guided mobile apps, Internet-based treatments), whereas the latter involves some level of therapist involvement (e.g., video teleconferencing, telephone support; Anton & Jones, 2017). Suffice to say that there is emerging evidence that family-based interventions delivered via the Internet, either as stand-alone programs (e.g., Sanders, Baker, & Turner, 2012), via videoconferencing to remotely deliver PMT (Comer et al., 2017), or as adjuncts to clinic-delivered interventions (e.g., Jones, Forehand, Cuellar, Parent, & Honeycutt, 2014) are effective with a variety of families of children with CP (see reviews by Breitenstein, Gross, & Christophersen, 2014; McGoron & Ondersma, 2015; Watson MacDonell & Prinz, 2017). In one study, an Internet version of PCIT (I-PCIT) provided stronger effects on some outcomes than therapist-delivered PCIT (Comer et al., 2017). Jones et al. (2014) presented preliminary evidence that a technology-enhanced version of HNC utilizing a smart phone app that included an HNC skills video series, brief daily surveys, text message reminders, video recording of home practice, and midweek video calls enhanced engagement and outcome, compared to HNC alone, for a sample of economically disadvantaged families. Researchers are now drawing attention to various challenges and issues involved in the uptake and implementation of technology-based interventions (e.g., Anton & Jones, 2017; Chou, Bry, & Comer, 2017), and Anton and Jones have provided a conceptual framework for facilitating uptake and implementation of technology-enhanced treatments by individual therapists as well as provider organizations. These novel approaches to the delivery of family-based interventions for youth CP hold promise for increasing the reach of such interventions to families (e.g., those in rural or under-resourced communities) who may not typically receive them.

Fidelity to treatment (i.e., the extent to which therapists adhere to the core components of a particular intervention) has a strong base of support showing that high fidelity to various evidence-based treatments, many of them described in this chapter, results in better outcomes than when therapists demonstrate poor fidelity to the treatment model (for reviews, see Garbacz, Brown, Spee, Polo, & Budd, 2014; Goense, Assink, Stams, Boendermaker, & Hoeve, 2016). GenerationPMTO and MST have been vanguards of this approach (e.g., Forgatch, Patterson, & DeGarmo, 2005; Henggeler & Schaeffer, 2017; Hukkelberg & Ogden, 2013). However, there is a pressing need for a standardized and comprehensive definition of fidelity that includes therapist adherence to the model, therapist competence (both with respect to the technical components of treatment as well as soft clinical skills), and treatment differentiation (Goense et al., 2016; Schoenwald et al., 2011). This then must be translated into reliable and valid measures of fidelity, and subsequent widespread adoption of fidelity assessment into clinical practice. The efforts by Forgatch and colleagues have been exemplary in this regard (e.g., Forgatch et al., 2005; Knutson, Forgatch, Rains, & Sigmarsdóttir, 2009).

A final direction for future research concerns recent developments in the translation of competing, or perhaps complementary, theoretical conceptualizations on the development of youth CP into novel family-based interventions. Historically, much of the empirical support on family-based treatments for child CP has been from interventions based on a social learning (or behavioral) model. This has been especially the case for PMT. There is some, but not uniform, support for the contention that social learning-based interventions are more effective than non-behavioral family-based interventions (for reviews, see Comer et al., 2013; Kaminski & Claussen, 2017), although as noted above, several recent individual studies have found comparable effects to social learning-based interventions for interventions based primarily on attachment theory (Högström et al., 2017), emotion coaching (Duncombe et al., 2016), and problem-solving (Ollendick et al., 2016). In addition, some evaluations of social learning-based treatments have documented improvements in attachment-related outcomes (e.g., maternal warmth, sensitivity) in addition to changes in parenting behaviors, such as praise and instruction giving (e.g., Blizzard, Barroso, Ramos, Graziano, & Bagner, 2017; O’Connor, Matias, Futh, Tantam, & Scott, 2013). Fisher and Skowron (2017) have recently suggested the compatibility of social learning and attachment perspectives for family-based interventions for a variety of child and family issues, and have noted that the field seems to be moving in the direction of “relational interventions” (p. 169). Such an approach might also incorporate more emotion-focused elements as well (e.g., Kaminski et al., 2008). In our own research, we are currently examining the feasibility of a combined intervention (HNC plus emotion coaching; McMahon et al., 2017) for clinic-referred children with ODD and CU traits.

Future Directions for Policy and Practice

In this section, we highlight four specific areas relevant to policy and practice: (a) the need to select evidence-based interventions; (b) family-based treatment as a core intervention for the treatment of youth CP; (c) family-based treatment as prevention; and (d) implementation in real-world settings.

Select Evidence-Based Interventions

Despite the available wealth of data pertaining to the outcomes of family-based interventions for youth CP, there is still a divide between clinical research and practice with respect to the implementation of empirically supported family-based programs. Considering the scarcity of resources in clinical care settings, along with clinicians’ ethical obligation to service clients according to best practice guidelines, it is critical that clinicians (and the policy-makers that fund such decisions) choose family-based treatment programs that have an adequate empirical base. There are many interventions (family-based and otherwise) that are available commercially that have anecdotal or practice-based evidence, but little or no empirical support. Yet these non-evidence-based programs are extensively used (Petrosino, MacDougall, Hollis-Peel, Fronius, & Guckenberg, 2015). Although these programs may prove to be effective in robust research trials, until these data are available, clinicians and policymakers should be encouraged to seriously consider this caveat. Reference to key reviews and meta-analyses (Kaminski & Claussen, 2017; McCart & Sheidow, 2016) and lists of evidence-based practices (e.g., California Evidence-based Clearinghouse for Child Welfare, 2017; Substance Abuse and Mental Health Services Administration, 2017) can be useful starting points for the identification of potential interventions.

Family-Based Treatment is a Core Intervention Component

There is overwhelming support for family-based treatment as an essential core intervention for reducing CP in youth. In fact, PMT may be sufficient as a stand-alone intervention for children with CP between the ages of 3–6 years. For older children and adolescents, multicomponent treatments that involve therapeutic work with the youth and his/her parents in the contexts of both the family and the broader community (e.g., school, peer group), are more often indicated. Nonetheless, family-based treatment should always be a core component in these multicomponent interventions.

Family-Based Treatment as Prevention

Traditionally, family-based interventions for youth CP have been considered to represent a form of treatment, rather than prevention. However, it is important to keep in mind that the boundaries between prevention and treatment are often very fluid. PMT treatment interventions for young children’s CP may have significant preventive effects (on the occurrence of later CP and delinquent behavior), especially if applied during the preschool years (e.g., Reid, 1993). An integrative review of 26 reviews and meta-analyses (1075 studies) of preventive interventions published between 1990 and 2008 found that PMT interventions had a larger effect size than either child-focused or school/community-based interventions (ds = 0.56, 0.41, and 0.28, respectively; Beelmann & Raabe, 2009). It can also be argued that family-based treatments for adolescents with CP also serve a preventative function, if they decrease the probability of entry into the justice system, or reduce the likelihood of future offending (see Salari and Enebrink (2018) for a detailed discussion of family-based preventive interventions).

Implementation in Real-World Settings

As noted above, there is a current emphasis on implementing family-based treatments in real-world settings (e.g., Gardner et al., 2016; Michelson et al., 2013). Thus, it is important to recognize the potential challenges faced by community stakeholders and intervention researchers in this collaborative endeavor. For example, referrals to community settings, such as child and family mental health centers, are often characterized by high rates of diagnostic comorbidity and case complexity, and difficult-to-engage families; furthermore, some isolated populations (e.g., rural families) cannot frequently access these services. Such obstacles call for creativity and innovation in remodeling the format and delivery of current family-based intervention approaches, while retaining the science underlying the intervention. In this light, some potential solutions to these challenges include those described in the previous section, including personalization of treatment, increased focus on the process of engagement, and the use of innovative adaptations of existing family-based treatments and/or delivery systems.

In addition to child- and family-informed barriers in implementation, other obstacles occur at the levels of individual providers or practitioners, and collaborating agencies (Southam-Gerow, Rodríguez, Chorpita, & Daleiden, 2012). For example, practitioners in community mental health services often differ in their levels of prior experience, education, and training in clinical work with families, which may facilitate or hinder the effectiveness of implementation efforts. Moreover, the organizational climate of an agency may involve high staff turnover, thereby reducing the number of available trained leaders and champions of an intervention to ensure successful implementation. Train-the-trainer models have been developed to help combat this obstacle by allowing agencies to adopt the necessary training resources to be self-sustaining in the ongoing implementation of family-based interventions (Dishion et al., 2016).

Conclusions

Family-based treatments are clearly the interventions of choice in treating child and adolescent CP. Research on these approaches has provided substantial empirical support for their efficacy, generalization, social validity, and effectiveness in a wide variety of settings and with various populations in the real world. Furthermore, there is increasing evidence for the benefits of family-based interventions from an economic perspective.

The evidence base for PMT interventions with younger children is relatively stronger than it is for family-based interventions with adolescents. This likely speaks to the entrenchment and increased variety and severity of CP behaviors in youth on the early-starter developmental pathway, as well as the broader set of contextual influences on the CP behavior (e.g., school, peer, and neighborhood) by the time that these youth become adolescents. However, it may also be partly due to the relative maturity of the empirical bases for these two types of intervention.

This relative difference in efficacy also speaks to the importance of viewing PMT with younger children as playing a key role not only in the treatment of children with CP, but as central to the prevention of later more serious antisocial behavior and criminal activity (see above).

Although family-based interventions have much to contribute to the treatment of children and adolescents with CP, they are clearly not a panacea. Too many children and families fail to respond sufficiently to these interventions—this must be a major focus of research and clinical practice moving forward. As noted above, there is much research activity focused on various aspects of this issue, and there is reason to be optimistic that the field will continue to advance. We owe this continued pursuit to the children and families who allow us (as clinicians, researchers, and policymakers) to enter their lives.