A key defining feature of family and multisystem approaches is that they treat adolescents in the context of the family and social systems in which substance use develops and may be maintained. Thus, inclusion of family members in treatment (often with the provision of home visits) is seen as a critical strategy for reducing attrition and addressing multiple issues simultaneously (Henggeler et al., 1996; Liddle et al., 2001). Because they are grounded solidly in the knowledge base on adolescents and development and thus are well suited to the specific problems of this population, family-based approaches have been among the most widely studied approaches for adolescents in controlled trials and have the highest levels of empirical support (Deas & Thomas, 2001; Liddle & Dakof, 1995; Waldron, 1997). Waldron et al. (2001) summarizes their success as follows:

Moreover, the high level of support for family and multisystemic approaches parallels findings from large meta-analyses pointing to the effectiveness of family therapies for adult substance users (Stanton & Shadish, 1997). It should be noted that family-based approaches are diverse, and many combine a variety of techniques, including family and individual therapies and skills and communication training, which may broaden the benefits of treatment by allowing greater individualization and enabling clinicians to address multiple factors in treatment (Waldron et al., 2001). Those family-based approaches with the highest level of support with this population include multisystemic therapy (MST) (Henggeler & Borduin, 1990), brief strategic family therapy (Szapocznik & Hervis, 2003), and multidimensional family therapy (Liddle et al., 2001).

Multisystemic therapy (MST) is a manualized approach that addresses the multiple determinants of drug use and antisocial behavior. It is intended to promote fuller family involvement through engaging family members as collaborators in treatment, stressing the strength of the youth and their families, and addressing a broad and comprehensive array of barriers to attaining treatment goals. Therapists must be familiar with several empirically based therapies (including structural family therapy and cognitive-behavioral therapy) and make frequent visits to the home and be available on a full-time basis to families. Henggeler and colleagues (1996) conducted a controlled trial with 118 substance-abusing or substance-dependent juvenile offenders (mean age 16) in which participants were randomly assigned to home-based MST and compared with usual community treatment services. The comparison condition involved referral by the youth's probation officer to outpatient adolescent group meetings. Ninety-eight percent of families completed a full course of treatment (an average of 130 days and 40 hr of service provision), compared with very little service access among the youth assigned to the control group (78% of youths received no substance abuse or mental health services, and only 5% received both substance use and mental health services). In other studies, MST has been shown to reduce re-arrest rates up to 64% and to be associated with significantly lower rates of substance-related arrests (Henggeler et al., 1991; Henggeler, Melton, Brondino, Scherer, & Hanley, 1997).

Brief strategic family therapy (BSFT; Szapocznik & Hervis, 2003) is a somewhat less intensive approach (as it targets fewer systems and can be delivered through a once-per-week office-

based format) that has also achieved an impressive level of empirical support. In BSFT, patterns of interaction in the family system are targeted that have been shown to influence adolescent drug abuse. The therapy consists of three classes of interventions: engaging all family members in treatment, identifying family strengths as well as roles and relationships linked to adolescent problems, and developing new family interactions (e.g., improved parenting skills and conflict resolution) to protect the adolescent. Home visits and use of specific engagement strategies are encouraged. In a study of 126 drug-abusing adolescents and their families that compared BSFT to a group control condition, 75% of those assigned to BSFT showed reliable improvement and 56% could be classified as recovered. In the control condition, only 14% showed reliable improvement, whereas 43% showed reliable deterioration in marijuana use (Santisteban et al., 2003). Brief strategic family therapy has also been shown to be associated with improved retention (Santisteban et al., 1996; Szapocznik et al., 1988) as well as significant reductions in the frequency of externalizing behaviors (aggression, delinquency) (Szapocznik, Kurtines, Foote, Perez-Vidal, & Hervis, 1986).

Multidimensional family therapy (MDFT) is a multicomponent, staged, family therapy that targets the substance-abusing adolescents, their families, and their interactions. Liddle and colleagues (2001) assigned 182 substance-abusing adolescents who were referred by the criminal justice system or the schools to either MDFT, group therapy, or multifamily education. Treatment was delivered in weekly sessions over 6 months, with roughly 70% of participants completing treatment across conditions. Superior outcomes for the adolescents assigned to MDFT relative to other approaches were seen at termination and 1-year follow-up. At termination, 42% of those assigned to MDFT, 25% of those in group therapy, and 32% of those in family education had clinically significant reductions in their drug use. Positive outcomes have also been reported for other models of family therapy, including family system therapy (Joanning, Thomas, & Quinn, 1992) and functional family therapy (Friedman, 1989).

A wide range of individual behavioral interventions, including those which seek to provide alternate reinforcers to drugs or reduce reinforcing aspects of abused substances, are based on operant conditioning theory and recognition of the reinforcing properties of abused substances (Aigner, 1978; Bigelow, Stitzer, & Liebson, 1984; Thompson & Pickens, 1971). Among adult substance users, these approaches have among the highest of empirical support (Griffith, Rowan-Szal, Roark, & Simpson, 2000; National Institute on Drug Abuse [NIDA], 2000). Examples include the work of Stitzer and colleagues, which has demonstrated that methadone-maintained opi-oid addicts will reduce illicit drug use when incentives such as take-home methadone are offered for abstinence (Stitzer & Bigelow, 1978; Stitzer, Iguchi, & Felch, 1992; Stitzer, Iguchi, Kidorf, & Bigelow, 1993). Contingency management incentive systems (Budney & Higgins, 1998; Budney, Higgins, Radonovich, & Novy, 2000; Higgins, Delany, Budney, Bickel, Hughes, et al., 1991, 1999; Kirby, Marlowe, Festinger, Lamb, & Platt, 1998; Petry, Martin, Cooney, & Kranzler, 2000; Silverman et al., 1996) offer incentives for targeted treatment goals (e.g., retention, drug-free urines) on an escalating schedule of reinforcement.

Behavioral approaches have only recently begun to be evaluated among substance-abusing adolescents. Azrin, Donahue, and Besalel (1994) assigned 26 substance-using adolescents to supportive counseling or behavior therapy, which consisted of therapist modeling and rehearsal, self-monitoring, and written assignments. After 6 months, urine toxicology screens and self-reports suggested significantly less substance abuse among the group assigned to behavioral therapy relative to supportive counseling, as well as better school and family functioning.

Contingency management approaches have not yet been widely used or evaluated with adolescents. In a feasibility study that involved adolescent smokers as a model for drug use, Corby, Roll, Ledgerwood, and Schuster (2000) found that providing cash incentives for not smoking to adolescents enrolled in a smoking cessation project (as assessed by twice daily CO levels) re

duced adolescent smoking and appeared to improve their mood. In a pilot study involving young adult marijuana users referred by the criminal justice system, Sinha, Easton, and Kemp (2003) studied the use of vouchers that could be used to purchase items in neighborhood stores. By providing these vouchers as rewards contingent on session attendance, treatment retention improved significantly.

Cognitive-behavioral approaches, based on social learning theory, are among the approaches with highest levels of empirical support for the treatment of adult substance use disorders. Key defining features of most cognitive-behavioral approaches for substance use disorders are (1) an emphasis on functional analysis of drug use, that is, understanding instances of substance use with respect to its antecedents and consequences, and (2) emphasis on skills training and self-regulation. Cognitive-behavioral therapy (CBT) has been shown to be effective across a wide range of substance use disorders (Carroll, 1996; Bowers, Dunn, & Wong, Irvin, 1999), including alcohol dependence (Miller & Wilbourne, 2002; Morgenstern & Longabaugh, 2000), marijuana dependence (MTP Research Group, 2001; Stephens, Roffman, & Curtin, 2000), cocaine dependence (Carroll, Rounsa-ville & Nich, 1994; Carroll, Nich, Ball, McCance-Katz, & Rounsaville, 1998; McKay, 1997; Rohsenow, Montl, Martin, Michalec, & Abrams, 2000), and nicotine dependence (Fiore, Smith, Jorenberg, & Baker, 1994; Hall, 1998; Patten et al., 1998). These findings are consistent with evidence supporting the effectiveness of CBT across a number of other psychiatric disorders as well, including depression, anxiety disorders, and eating disorders (DeRubeis & Crits-Christoph, 1998).

Cognitive-behavioral therapy has also been evaluated as a treatment for adolescent substance use disorders. In an extremely well-done study, Waldron and colleagues (2001) randomly assigned 120 adolescents who were abusers of illicit drugs (primarily marijuana) to one of four treatment conditions: family therapy alone (functional family therapy), individual CBT alone, a combination of individual and family therapy, and a psychoeducational group. Completion rates were high (70% to 80% across groups). In general, while there were meaningful reductions in drug use in all conditions, there were larger and more durable reductions in substance use for the combined and family conditions relative to the individual CBT and group conditions. Treatment effects were strongest immediately after treatment but persisted through a 7-month follow-up.

Kaminer and colleagues (2002) compared group CBT to psychoeducational substance abuse treatment for 88 adolescents referred for treatment of a substance abuse problem. Eighty-six percent of the sample completed treatment and 9-month follow-up data were available for 65% of the sample. The presence of a conduct disorder was associated with treatment dropout. Cognitive-behavioral therapy was significantly more effective than psychoeducation only for male subjects; females appeared to improve regardless of treatment condition. Nevertheless, there were no significant differences between the two conditions at the 9-month follow-up. The relatively high rates of relapse in this sample (52% had a urinalysis that was positive for marijuana at the 9-month follow-up evaluation) suggest that an eight-session stand-alone approach may not be adequately intensive or structured for this population.

Motivational approaches are brief treatment approaches designed to produce rapid, internally motivated change in addictive behavior and other problem behaviors. Grounded in principles of motivational psychology and patient-centered counseling, motivational interviewing (MI; Miller & Rollnick, 1991, 2002) arose out of several recent theoretical and empirical advances (Miller, 2000). Motivational interviewing has a high level of empirical support in the adult substance abuse treatment literature (Burke, Arkowitz, & Menchola, 2003; Dunn, Deroo, & Rivara, 2001; Miller, 2002; Wilk, Jensen, & Havighurst, 1997). The core principles of MI are as

follows: (1) express empathy; (2) develop discrepancy; (3) avoid argumentation; (4) roll with resistance; (5) support self-efficacy. Motivational interviewing makes the important assumption that ambivalence and fluctuating motivations define substance abuse recovery and need to be thoroughly explored rather than confronted harshly. Ambivalence is considered a normal event, not something that indicates the patient is unsuitable for treatment or needs vigorous confrontation in hopes of forcing a sudden change. The patient's point of view is respected, which in some cases may mean accepting that major change, or even any change, is not what the patient wants, at least at the present time (Carroll, Ball, & Martino, 2004). Thus, while the bulk of research on efficacy of MI is in the adult literature, this nonconfrontational approach appears quite well suited for application to adolescents and young adults, given its flexibility around goals and recognition of abstinence as part of the change process.

Another distinct advantage of using MI with adolescent populations is that it can be implemented in a range of settings, given that adolescents with substance abuse problems rarely seek treatment of their own volition in traditional substance abuse settings. Monti and colleagues (1999) studied 94 adolescents treated at an emergency room for a problem related to alcohol use (e.g., injuries related to drinking, drunk driving). They were randomly assigned to MI or standard care, with all interventions and assessments conducted in the emergency room. At a 6-month follow-up, there were significantly fewer incidents of drunk driving, traffic violations, and alcohol-related problems in the group assigned to receive MI. Not only does this study suggest the promise of brief motivational approaches for this population, but it also underlines the importance of intervening with adolescents in nontraditional settings.

While disease model treatments and other approaches associated with the Twelve Steps of Alcoholics Anonymous dominate the treatment system for both adults and adolescents, there are no randomized controlled trials evaluating the effectiveness of these approaches in adolescents. Recent reports from randomized controlled trials evaluating the efficacy of manualized Twelve-Step approaches have found evidence to suggest their effectiveness with adult substance users (Carroll et al., 1998; Crits-Christoph et al., 1999; Project MATCH Research Group, 1997). It is important to note, however, that these manual-guided approaches are highly structured, delivered as individual (rather than group) therapy, and might be quite different from the nonmanualized group approaches typically delivered in community settings with adolescents. In addition, since individual drug counseling emphasizes and encourages frequent Twelve-Step group attendance, its effectiveness might reflect increased patient involvement in rehabilitative groups. It is important to note that the absence of sufficient research on Twelve-Step treatment should not lead one to conclude that this widespread and popular approach is ineffective.

Data on the effectiveness of more traditional programs are beginning to emerge, but no data from randomized trials comparing these approaches to alternatives are available. Winters and colleagues (2000) reported on a large nonrandomized evaluation comparing a group of substance-abusing youth who completed the Twelve-Step Minnesota Model treatment to similar individuals who did not complete treatment and to a group on a waiting list for treatment. The treatment was multimodal, based on the principles of the Twelve Steps of Alcoholics Anonymous, and included group therapy and individual counseling, family therapy, lectures about the Twelve Steps, and reading assignments. Better substance use and psychosocial outcomes at 6 and 12 months were reported for those who completed treatment compared with those who did not complete or who did not receive treatment. Although a high rate of abstinence was reported among treatment completers, it is difficult to interpret these findings, given the self-selection due to lack of randomization and lack of measurement of treatment delivery or process.

Recently, the Center for Substance Abuse Treatment (CSAT) funded the largest multisite clinical trial comparing an array of diverse outpatient treatments targeting adolescent marijuana abuse and dependence. The study, conducted at four sites, involved 600 randomized patients and evaluated five manualized treatments. Notably, the study had a follow-up rate of 95% for up to 30 months. Treatment modalities covered the full range of treatments (individual, group, family, and comprehensive multicomponent) that took place between 5 and 12 weeks and included 6 to 21 sessions (Diamond et al., 2002). All five treatments performed equally well and were associated with marked (50%) reductions in frequency of marijuana use; these improvements were maintained through a 30-month follow-up evaluation. It is noteworthy that these very promising outcomes were seen even for the less costly 6-week, five-session treatment.

As new effective therapies for adolescents are identified, it is imperative that the field move toward evaluating how these treatments exert their effects, by looking at mediators and moderators of outcome. Several recent investigations have examined these variables. In terms of retention and engagement, Szapocznik and colleagues' impressive work on engaging teens and families in treatment has been replicated and further developed (e.g., Coatsworth, Santisteban, McBride, & Szapocznik, 2001; Santisteban et al., 1996). Henggeler and colleagues (1996) demonstrated a 98% treatment completion rate for home-based MST (Szapocznik et al., 1983). They have also demonstrated that adherence to the treatment was significantly associated with better treatment outcome. Liddle and colleagues have conducted several process studies looking at mechanisms of change, including in-session patterns of change associated with the resolution of parent–adolescent conflict (Diamond & Liddle, 1996) and the link between improvement in parenting and better substance use outcome (Schmidt, Liddle, & Dakof, 1996). These kinds of studies will help identify key treatment ingredients that might lead to increased treatment potency.

Another area relevant to treatment research that may be influenced by developmental perspective concerns assessment. While self-report of substance use by adolescents has been confirmed as fairly reliable (Buchan, Dennis, Tims, & Diamond, 2002), recent analysis from the Cannabis Youth Treatment Study suggests the addition of parent reports adds additional information not provided by the adolescent. Although adolescents and parents reported about the same number of substance use symptoms, there was a very low concordance between the types of symptom endorsed. Parents tended to report more symptoms related to role failure, tolerance, and substance-induced psychological problems (Dennis, Babor, Roebuck, & Donaldson, 2002). Similar findings were discovered regarding mental health symptoms. Specifically, parents tended to endorse more symptoms of depression and attention problems (Diamond, Panichelli-Mindel, Shera, Tims, & Ungemack, in press). This was particularly true for African-American adolescents. Thus parent report may have a unique contribution when working with a minority population, a community that has been characterized as suspicious of the research community.

One area that has received strikingly little research with adolescents is the integration of substance use and other mental health services that can treat adolescents with both kinds of disorders. Historically there has been a divide between treatment systems for substance abuse and mental health disorders. Substance abuse counselors often have little or no training in mental health issues, and programs either ignore co-occurring problems or refer patients to other systems during (parallel) or after (sequential)