Chapter 6 - Clinical Research Supporting Brief Stategic Family Therapy

This chapter describes past research on the effectiveness of BSFT with drug-abusing adolescents with behavioral problems. BSFT has been found to be effective in reducing adolescents' conduct problems, drug use, and association with antisocial peers and in improving family functioning. In addition, BSFT engagement has been found to increase engagement and retention in therapy. Additional studies testing an ecological version of BSFT with this population are currently underway.

As presented in this manual, BSFT's primary emphasis is on identifying and modifying maladaptive patterns of family interaction that are linked to the adolescent's symptoms. The ecological version of BSFT, BSFT-ecological (Robbins et al. in press) applies this principle of identifying and modifying maladaptive patterns of interaction to the multiple social contexts in which the adolescent is embedded (cf. Bronfenbrenner 1979). The principal social contexts that are targeted in BSFT-ecological are family, family-peer relations, family-school relations, family-juvenile justice relations, and parent support systems. Joining, diagnosing, and restructuring, as developed in BSFT to use within the family system, are applied to these other social contexts or systems that influence the adolescent's behaviors. For instance, the BSFT counselor assesses the maladaptive, repetitive patterns of interaction that occur in each of these systems or domains. As an example, the BSFT counselor would diagnose the family-school system in the same way that he or she would diagnose the family system. In diagnosing structure, the counselor would ask, "Do parents provide effective leadership in their relationship with their child's teachers?" In diagnosing resonance, the counselor would ask, "Are parents and teachers disengaged?" In diagnosing conflict resolution, the counselor's questions would be, "What is the conflict resolution style in the parentteacher relationship? Might parents and teachers avoid conflict with each other (by remaining disengaged) or diffuse conflicts by blaming each other?" In BSFT-ecological, joining the teacher in the parentteacher relationship employs the same joining techniques developed for BSFT. Similarly, in BSFT-ecological, BSFT restructuring techniques are used to modify the nature of the relationship between a parent and his or her child's teacher.

A recent study (Santisteban et al. in press) examined the efficacy of BSFT in reducing an adolescent's behavioral problems, association with antisocial peers, and marijuana use, and in improving family functioning. In this study, outpatient BSFT was compared to an outpatient group counseling control treatment. Participants were 79 Hispanic families with a 12- to 18-year-old adolescent who was referred to counseling for conduct and antisocial problems by either a school counselor or a parent. Families were randomly assigned to either BSFT or group counseling. Analyses of treatment integrity revealed that interventions in both therapies adhered to treatment guidelines and that the two therapies were clearly distinguishable.

Conduct disorder and association with antisocial peers were assessed using the Revised Behavior Problem Checklist (RBPC) (Quay and Peterson 1987), which is a measure of adolescent behavior problems reported by parents. Conduct disorder was measured using 22 items, and association with antisocial peers was measured using 17 items. Each item asks the parent(s) to rate whether a specific aspect of the adolescent's behavior (e.g., fighting, spending time with "bad" friends) is no problem (0), a mild problem (1), or a severe problem (2). Ratings for all items on each scale are then added together to derive a total score.

The effects of BSFT on conduct disorder, association with antisocial peers, and marijuana use were evaluated in two ways. First, analyses of variance were conducted to examine whether BSFT reduced conduct disorder, association with antisocial peers, and marijuana use to a significantly greater extent than did group counseling. Second, exploratory analyses were conducted on clinically significant changes in conduct problems and association with antisocial peers. These exploratory analyses used the twofold clinical significance criteria recommended by Jacobson and Truax (1991). To be able to classify a change in symptoms for a given participant as clinically significant, two conditions have to occur. First, the magnitude of the change must be large enough to be reliable--that is, to rule out random fluctuation as a plausible explanation. Second, the participant must "recover" from clinical to nonclinical levels, i.e., cross the diagnostic threshold.

Conduct Disorder. Analyses of variance indicated that conduct disorder scores for adolescents in BSFT compared to those for adolescents in group counseling were significantly reduced between pre- and posttreatment. In the clinical significance analyses, a substantially larger proportion of adolescents in BSFT than in group counseling demonstrated clinically significant improvement. At intake, 70 percent of adolescents in BSFT had conduct disorder scores that were above clinical cutoffs. That is, they scored above the empirically established threshold for clinical diagnoses of conduct disorder. At the end of treatment, 46 percent of these adolescents showed reliable improvement, and 5 percent showed reliable deterioration. Among the 46 percent who showed reliable improvement, 59 percent recovered to nonclinical levels of conduct disorder. In contrast, at intake, 64 percent of adolescents in group counseling had conduct disorder scores above the clinical cutoff. Of these, none showed reliable improvement, and 11 percent showed reliable deterioration. Therefore, while adolescents in BSFT who entered treatment at clinical levels of conduct disorder had a 66 percent likelihood of improving, none of the adolescents in group counseling reliably improved.

Association With Antisocial Peers. Analyses of variance indicated that, for adolescents in BSFT, scores for association with antisocial peers were significantly reduced between pre- and post-treatment, compared to those for adolescents in group counseling. In the clinical significance analyses, 79 percent of adolescents in BSFT were above clinical cutoffs for association with antisocial peers at intake. Among adolescents in BSFT meeting clinical criteria for association with antisocial peers, 36 percent showed reliable improvement, and 2 percent showed reliable deterioration. Of the 36 percent of adolescents in BSFT with reliable improvement, 50 percent were classified as recovered. Among adolescents in group counseling, 64 percent were above clinical cutoffs for association with antisocial peers at intake. Among adolescents in group counseling meeting these clinical criteria at intake, 11 percent reliably improved, and none reliably deteriorated. Of the 11 percent of adolescents in group counseling evidencing reliable improvement in association with antisocial peers, 50 percent recovered to nonclinical levels. Hence, adolescents in BSFT who entered treatment at clinical levels of association with antisocial peers were 2.5 times more likely to reliable improve than were adolescents in group treatment. Marijuana Use. Analyses of variance revealed that BSFT was associated with significantly greater reductions in self-reported marijuana use than was group counseling. To investigate whether clinically meaningful 3 changes in marijuana use occurred, four use categories from the substance use literature (e.g., Brooks et al.1998) were employed. These categories are based on the number of days an individual uses marijuana in the 30 days before the intake and termination assessments:

• abstainer - 0 days
• weekly user - 1 to 8 days
• frequent user - 9 to 16 days
• daily user - 17 or more days

In BSFT, 40 percent of participants reported using marijuana at intake and/or termination. Of these, 25 percent did not show change, 60 percent showed improvement in drug use, and 15 percent showed deterioration. Of the individuals in BSFT who shifted into less severe categories, 75 percent were no longer using marijuana at termination. In group counseling, 26 percent of participants reported using marijuana at intake and/or termination. Of these, 33 percent showed no change, 17 percent showed improvement, and 50 percent deteriorated. The 17 percent of adolescents in group counseling cases that showed improvement were no longer using marijuana at termination. Hence, adolescents in BSFT were 3.5 times more likely than were adolescents in group counseling to show improvement in marijuana use.

Treatments also were compared in terms of their influence on family functioning. Family functioning was measured using the Structural Family Systems Ratings (Szapocznik et al. 1991). This measure was constructed to assess family functioning as defined in Chapter 3. Based on their scores when they entered therapy, families were separated by a median split into those who had good and those who had poor family functioning. Within each group (i.e., those with good and those with poor family functioning), a statistical test that compares group means (analysis of variance) tested changes in family functioning from before to after the intervention.

Among families who were admitted with poor family functioning, the results showed that those assigned to BSFT had a significant improvement in family functioning, while those families assigned to group counseling did not improve significantly.

Among families who were admitted with good family functioning, the results showed that those assigned to BSFT retained their good levels of family functioning, while families assigned to group counseling showed significant deterioration. These findings suggest that not all families of drug-abusing youths begin counseling with poor family functioning, but if the family is not given adequate help to cope with the youth's problems, the family's functioning may deteriorate.

With the advent of the adolescent drug epidemic of the 1970s, the vast majority of counselors who worked with drug-using youths reported that, although they preferred to use family therapy, they were not able to bring whole families into treatment (Coleman and Davis 1978). In response, a procedure was developed that would achieve the goals of BSFT (to change maladaptive family interactions and symptomatic adolescent behavior) without requiring the whole family to attend treatment sessions. The procedure is an adaptation of BSFT called "One Person" BSFT (Szapocznik et al. 1985; Szapocznik and Kurtines 1989; Szapocznik et al. 1989a). One Person BSFT capitalizes on the systemic concept of complementarity, which suggests that when one family member changes, the rest of the system responds by either restoring the family process to its old ways or adapting to the new changes (Minuchin and Fishman 1981). The goal of One Person BSFT is to change the drug-abusing adolescent's participation in maladaptive family interactions that include him or her. Occasionally, these changes create a family crisis as the family attempts to return to its old ways. The counselor uses the opportunity created by these crises to engage reluctant family members.

A clinical trial was conducted to compare the efficacy of One Person BSFT to Conjoint (full family) BSFT (Szapocznik et al. 1983, 1986). Hispanic families with a drug-abusing 12- to 17-year-old adolescent were randomly assigned to the One Person or Conjoint BSFT modalities. Both therapies were designed to use exactly the same BSFT theory so that only one variable (one person vs. conjoint meetings) would differ between the treatments. Analyses of treatment integrity revealed that interventions in both therapies adhered to guidelines and that the two therapies were clearly distinguishable. The results showed that One Person was as efficacious as Conjoint BSFT in significantly reducing adolescent drug use and behavior problems as well as in improving family functioning at the end of therapy. These results were maintained at the 6-month followup (Szapocznik et al. 1983, 1986).

One Person BSFT is not discussed in this manual because it is considered a very advanced clinical technique. More information on One Person BSFT is available in Szapocznik and Kurtines (1989).

As discussed in Chapter 5, in response to the problem of engaging resistant families, a set of engagement procedures based on BSFT principles was developed (Szapocznik and Kurtines 1989; Szapocznik et al. 1989b). These procedures are based on the premise that resistance to entering treatment can be understood in family interactional terms.

One Person BSFT techniques are useful in this initial phase. That's because the person who contacts the counselor to request help may become the one person through whom work is initially done to restructure the maladaptive family interactions that are maintaining the symptom of resistance. The success of the engagement process is measured by the family's and the symptomatic youth's attendance in family therapy. In part, success in engagement permits the counselor to redefine the problem as a family problem in which all family members have something to gain. Once the family is engaged in treatment, the focus of the intervention is shifted from engagement to removing the adolescent's presenting symptoms.

The efficacy of BSFT engagement has been tested in three studies with Hispanic youths (Szapocznik et al. 1988; Santisteban et al. 1996; Coatsworth et al. 2001). The first study (Szapocznik et al. 1988) included mostly Cuban families with adolescents who had behavior problems and who were suspected of or observed using drugs by their parents or school counselors. Of those engaged, 93 percent actually reported drug use. Families were randomly assigned to one of two therapies: BSFT engagement or engagement as usual (the control therapy). The engagement-as-usual therapy consisted of the typical engagement methods used by community treatment agencies, which were identified prior to the study using a community survey of outpatient agencies serving drug-abusing adolescents. All families who were successfully engaged received BSFT. In the experimental therapy, families were engaged and retained using BSFT engagement techniques. Successful engagement was defined as the conjoint family (minimally the identified patient and his or her parents and siblings living in the same household) attending the first BSFT session, which was usually to assess the drug-using adolescent and his or her family. Treatment integrity analyses revealed that interventions in both engagement therapies adhered to prescribed guidelines using six levels of engagement effort that were operationally defined and that the therapies were clearly distinguishable by level of engagement effort applied.

The six levels of engagement effort, as enumerated in Szapocznik et al. (1988, p. 554), are:

• Level 0 - expressing polite concern, scheduling an intake appointment, establishing that cases met criteria for inclusion in the study, and making clear who must attend the intake assessment;
• Level 1 - attempting minimal joining, encouraging the caller to involve the family, asking about the depth and breadth of adolescent problems, and asking about family members;
• Level 2 - attempting more thorough joining; asking about family interactions; seeking information about the problems, values, and interests of family members; supporting and establishing an alliance with the caller; beginning to establish leadership; and asking whether all family members would be willing to attend the intake appointment;
• Level 3 - restructuring for engagement through the caller, advising the caller about negotiating and reframing, and following up with family members (either over the phone or personally with the caller at the therapist's office) to be sure that intake appointments would be kept;
• Level 4 - conducting lower level ecological engagement interventions, joining family members or conducting intrapersonal restructuring (with family members other than the original caller) over the phone or in the therapist's office, and contacting significant others (by phone) to gather more information; and
• Level 5 - conducting higher level ecological interventions, making out-of-office visits to family members or significant others, and using significant others to help conduct restructuring.

Level 0-1 behaviors were permitted for both the BSFT engagement and engagement-as-usual conditions. Level 2-5 behaviors were permitted only for the BSFT engagement condition. Efficacy was measured in rates of both family treatment entry as well as retention to treatment completion.

The efficacy of the two methods of engagement was measured by the percentage of families who entered treatment and the percentage of families who completed the treatment. The results revealed that 42 percent of the families in the engagement-as-usual therapy and 93 percent of the families in the BSFT engagement therapy were successfully engaged. In addition, 25 percent of engaged cases in the engagement-as-usual treatment and 77 percent of engaged cases in the BSFT engagement treatment successfully completed treatment. These differences in engagement and retention between the two methods of engagement were both statistically significant. Improvements in adolescent symptoms occurred but were not significantly different between the two methods of engagement. Thus, the critical distinction between the treatments was in their different rates of engagement and retention. Therefore, BSFT engagement had a positive impact on more families than did engagement as usual.

In addition to replicating the previous engagement study, the second study (Santisteban et al. 1996) also explored factors that might moderate the efficacy of the engagement interventions. In contrast to the previous engagement study, Santisteban et al. (1996) more stringently defined the success of engagement as a minimum of two office visits: the intake session and the first therapy session. The researchers randomly assigned 193 Hispanic families to one experimental and two control treatments. The experimental therapy was BSFT plus BSFT engagement. The first control therapy was BSFT plus engagement as usual, and the second was group counseling plus engagement as usual. In both control treatments, engagement as usual involved no specialized engagement strategies.

Results showed that 81 percent of families were successfully engaged in the BSFT plus BSFT engagement experimental treatment. In contrast, 60 percent of the families in the two control therapies were successfully engaged. These differences in engagement were statistically significant. However, the efficacy of the experimental therapy procedures was moderated by the cultural/ethnic identity of the Hispanic families in the study. Among families assigned to BSFT engagement, 93 percent of the non-Cuban Hispanics (composed primarily of Nicaraguan, Colombian, Puerto Rican, Peruvian, and Mexican families) and 64 percent of the Cuban Hispanics were engaged. These findings have led to further study of the mechanism by which culture/ethnicity and other contextual factors may influence clinical processes related to engagement (Santisteban et al. 1996; Santisteban et al. in press). The results of the Szapocznik et al. (1988) and Santisteban et al. (1996) studies strongly support the efficacy of BSFT engagement. Further, the second study with its focus on cultural/ethnic identity supports the widely held belief that therapeutic interactions must be responsive to contextual changes in the treatment population (Sue et al. 1994; Szapocznik and Kurtines 1993).

A third study (Coatsworth et al. 2001) compared BSFT to a community control intervention in terms of its ability to engage and retain adolescents and their families in treatment. An important aspect of this study was that an outside treatment agency administered the control intervention. Because of that, the control intervention (e.g., usual engagement strategies) was less subject to the influence of the investigators. Findings in this study, as in previous studies, showed that BSFT was significantly more successful, at 81 percent, in engaging adolescents and their families in treatment than was the community control treatment, at 61 percent. Likewise, among those engaged in treatment, a higher percentage of adolescents and their families in BSFT, at 71 percent, were retained in treatment compared to those in the community control intervention, at 42 percent. In BSFT, 58 percent of adolescents and their families completed treatment compared to 25 percent of those in the community control intervention. Families in BSFT were 2.3 times more likely both to be engaged and retained in treatment than were families randomized to the community control treatment.

An additional finding of the Coatsworth et al. (2001) study warrants special mention. In BSFT, families of adolescents with more severe conduct problem symptoms were more likely to remain in treatment than were families of adolescents whose conduct problem symptoms were less severe. The opposite pattern was evident in the community control intervention, with families that were retained in treatment showing lower intake levels of conduct problems than did families who dropped out. These findings are particularly important because they suggest that adolescents who are most in need of services are more likely to stay in BSFT than in traditional community treatments.

References