Family-based treatments for adolescent drug abuse and related behavior problems have been developed and evaluated with success. Empirical support exists for the efficacy of family-based treatments, and process studies have begun to identify mechanisms by which these treatments may achieve their effects. This article discusses theory and related clinical refinements in a contemporary family-based intervention, multidimensional family therapy. Expansions in the theoretical basis of the model are discussed. I highlight 2 aspects of the theory evolution process, resulting in a sharper clinical focus on intrapersonal development and on adolescents' and families `functioning vis-a-vis influential extra familial ecologies of development.

MDFT is an outpatient, family-based treatment developed for multiproblem adolescents, in particular those teenagers presenting with drug and behavior problems (Liddle, 1992). The approach strives for a consistent and clinically practical connection among its organizational levels: theory, principles of intervention, and intervention strategies and methods. The interventions of the model derive from overarching principles of MDFT theory and target population characteristics, and they are guided by research-based knowledge about dysfunctional and normative adolescent and family development (Liddle, 1998). Interventions target the multiple ecologies of adolescent development and within these ecologies, they address the circumstances and processes known to produce or to continue dysfunction (Liddle, Rowe, Dakof, & Lyke, 1998). Although the same developmental challenges may be common to all adolescents and their families, we work to understand the differential and idiosyncratic individual adolescent and family expressions of these generic developmental challenges. Consistent with the recommendations of those who have discussed the need for a strong developmental basis in child and adolescent therapy (Vernberg, Routh, & Koocher, 1992), we strive to understand the unique manifestations of developmental competence and detours with each case.

The approach has been operationalized into different treatment applications. Versions of this manualized model have varied on dimensions such as treatment length, dosage and intensity, intervention locale (i.e., in-clinic or combination in-clinic and home based), and inclusion of particular therapeutic methods, such as the clinical use of within-treatment drug screens and intensive case management for the family (Liddle, 1998). The model has been developed and tested since 1985 in randomized clinical trials and treatment development and process studies (Liddle & Hogue, in press-b). These studies were conducted at different locations in the United States, including Philadelphia, the San Francisco Bay area, and Miami. The study populations were ethnically diverse, and their problem severity varied as well. Study participants have included high-risk early adolescents (Liddle & Hogue, in press-a) and multi-problem, juvenile-justice-involved, dually diagnosed, female and male adolescent substance abusers (Liddle & Dakof, 1995). This approach has been recognized as part of a new generation of comprehensive, multi-component, theoretically derived and empirically supported adolescent drag abuse treatments (Center for Substance Abuse Treatment, 1999; Lebow & Gurman, 1995; Nichols & Schwartz, 1998; Robbins, Szapocznik, Alexander, & Miller, in press; Selekman & Todd, 1990; Stanton & Shadish, 1997; Waldron, 1997; Weinberg, Rahdert, Colliver, & Glantz, 1998; Winters, Latimer, & Stinchfield, 1999).

From Structural-Strategic to MDFT
First-generation family therapy models differentiated themselves from individual and group therapies in a variety of ways. Family therapy, in the classic sense of the term, invoked a unit of analysis and intervention that honed in on the family per se. Individual and intrapersonal processes or extrafamilial processes were not emphasized in early era family therapy (Nichols & Schwartz, 1998). New family therapy models developed over time, however. The theoretical boundaries of new family intervention approaches were more comprehensive and were not based only in family systems theory. Today, in the adolescent treatment specialty, a few new treatments exemplify these changes. These approaches are functional family therapy (Alexander & Parsons, 1982), multisystemic therapy (Henggeler & Borduin, 1990; Henggeler et al., 1997), MDFT (Liddle, 1992, 1998), and structural ecosystems therapy (Szapocznik & Coatsworth, in press). In MDFT, for instance, the intrapersonal psychosocial functioning of several individuals in a family is understood as complementary to and interdependent with theoretical tenets that were the stock in trade of family therapy--a focus on intrafamilial behavioral interactions. A second major development in family therapy theory concerns how circumstances and events outside of the family--the role of ecological factors--are understood as influential in the promotion of development as well as the creation of dysfunction (Bronfenbrenner, 1979). The first version of MDFT was called structural-strategic family therapy (Liddle, 1984, 1985). This early version of the model aimed to be integrative. It blended the in-session change strategies of structural family therapy, with its emphasis on changing the in-session patterns of family interaction through enactment (S. Minuchin, 1974), and problem-solving therapy (Haley, 1976), with its focus on changing out-of-session sequences of behavior via behavioral tasks.

Theory of Dysfunction
Although less obvious than in-session therapy techniques, presumptions about how dysfunction develops, maintains, or worsens are implicit in every intervention and in the overall model in which the interventions reside. Ideally, the techniques of a treatment derive from the premises of a model about the development and persistence of dysfunction. Interventions, actualized via particular techniques, target the phenomena of interest (e.g., session content, person characteristics, and interpersonal processes). Treatment models should also specify processes or mechanisms by which the therapy techniques affect the intervention targets. Specification should include statements about how interventions influence the relevant domains of functioning.

Theory of Change
Therapy techniques are the agency-oriented extensions of the premises and knowledge--the theor