A Relational Frame Analysis of Defusion Interactions in Acceptance and Commitment Therapy: Preliminary and Quasi‑Experimental Study with At‑Risk Adolescents

Full reference: Luciano, C., Ruiz, F. J., Vizcaíno Torres, R. M., Sánchez Martín, V., Gutiérrez Martínez, O., & López López, J. C. (2011). A Relational Frame Analysis of Defusion Interactions in Acceptance and Commitment Therapy. A Preliminary and Quasi-Experimental Study with At-Risk Adolescents. International Journal of Psychology and Psychological Therapy, 11(2), 165-182.

Study type: Quasi-experimental study with between-subject and within-subject comparisons

Background and objectives

The present study aims to analyze the interactions involved in some of the defusion exercises typical of Acceptance and Commitment Therapy (ACT) according to an analysis based on Relational Frame Theory (RFT). Defusion exercises in ACT are oriented toward promoting discrimination of the ongoing process of having any thought or feeling as well as discriminating the person from who is having each of those thoughts. According to RFT, these interactions can be differentiated in terms of the transformations of functions involved (Type A, B, C, and D interactions).

The objective was to conduct a preliminary and quasi-experimental analysis of two defusion protocols that were designed with the mentioned interactions. Two protocols were compared: Defusion I (built with Type B interactions—deictic frames) and Defusion II (which included Type B interactions plus Type C interactions—hierarchical frames with examples—and Type D interactions—behavior-regulating function from a values perspective). The hypothesis was that both protocols would produce changes in measured variables, but Defusion II would produce more relevant changes due to its more complex structure with multiple types of interactions.

Method

Participants

Fifteen adolescents (12-15 years old, M = 13.66, SD = 0.9) participated in the study. All were students at a local secondary school. Seven were in grade 1, four in grade 2 respectively, and four in grade 3. All participants showed high scores (one standard deviation above the mean) on either the impulsivity subscale or the emotional symptoms subscales of the Behavior Assessment System for Children (BASC; Reynolds & Kamphaus, 2004).

Participants were categorized as high-risk (6 or more problematic behaviors, n = 6) or low-risk (fewer than 6 problematic behaviors, n = 9) based on the number of problematic behaviors reported by participants during an initial assessment of risk behaviors. Problematic behaviors were operationalized according to the Impulsive Behavior Inventory (IBI) and Emotional Behavior Inventory (EBI), which include lists of behaviors occurring in different contexts.

Experimental design

A quasi-experimental design with repeated measures was used with between-subject comparisons (Defusion I Protocol vs. Defusion II Protocol) and within-subject comparisons. Participants were distributed into three tutorial classes according to school organization. Although participants differed in the number and type of problematic behaviors, other factors of school organization allowed group organization. Consequently, one protocol was implemented in two groups while the other in only one.

In the first tutorial session, all three groups received the values-oriented session and then all participants were assessed. Subsequently, participants in grade 2 (aged 13-14) received Defusion I protocol while those in grade 1 (aged 12-13) and grade 3 (aged 14-15) received Defusion II protocol. The same protocol structure was implemented weekly with the three grades, with one different session each week for each class.

Materials and instruments

The following self-report questionnaires were used:

Behavior Assessment System for Children (BASC): Evaluated adaptive and non-adaptive behaviors (impulsivity, somatization, anxiety, depression) through participant and teacher reports.

Spanish Avoidance and Fusion Questionnaire (AFQ-S): A 9-item questionnaire constructed with 5 items from the Avoidance and Fusion Questionnaire (AFQ-Y; Greco, Lambert, & Baer, 2008) and 4 items from the Willingness and Acceptance Measure (WAM; Greco, Murrell, & Coyne, 2004). The AFQ-Y is a measure of psychological inflexibility (0-4). The WAM is a 14-item measure of experiential acceptance emphasizing children's ability to take action and engage in meaningful activities when experiencing private events. Scores on AFQ-S and KIMS were converted to a scale of 0 to 10.

Accepting without Judgment Scale of the Kentucky Inventory of Mindfulness Skills (KIMS): A brief 4-item version of the KIMS was used based on the good internal consistency (α = .84) found in a previous study (Ruiz, Langer, Luciano, Cangas, & Beltrán, under review).

Impulsive Behavior Inventory (IBI) and Emotional Behavior Inventory (EBI): Designed specifically for the study. They consisted of two lists of behaviors occurring in different contexts (28 items referring to impulsive reactions and 28 items referring to depressive or anxious reactions respectively).

Procedure

The research was designed as preliminary and quasi-experimental analysis. In the first session (pre-intervention) after responding to questionnaires on psychological inflexibility (AFQ-S), problematic behaviors (IBI or EBI) and emotional symptom indices from the BASC, a values-oriented session was conducted. Two weeks later, 9 low-risk participants received either Defusion I protocol (n = 4) or Defusion II protocol (n = 5). Six high-risk participants (with more than 6 problematic behaviors) received Defusion II protocol.

Defusion I Protocol: Was constructed on the basis of deictic framing trials (Type-B triads) that allowed participants to notice the ongoing process of having any thought or feeling, as well as to notice deictic behaviors to discriminate the individual from the ongoing process of having thoughts/feelings. The protocol included at least 14 exercises implemented in an experiential-computerized format.

Defusion II Protocol: Was constructed with Type B interactions (deictic frames) plus Type C interactions (hierarchical frames with examples) and Type D interactions (questions to promote behavior-regulating function from a values perspective). The Defusion II protocol included all types of interactions (B, C, D) in Defusion I but also included iterative questions that differentiated what was added in Defusion II—namely, questions to promote behavior-regulating function from a consistent perspective (choosing what to do, from I-Here, with impulses or emotional problems and with thoughts about what is important, all in I-There).

Two weeks after protocol implementation, a post-defusion assessment was conducted with the same questionnaires (AFQ-S, IBI or EBI, KIMS). Four months later, a follow-up assessment was conducted.

Data analysis

Within-subject data analyses used Wilcoxon Z and Mann-Whitney U tests for between-condition comparisons. Nonparametric tests were applied due to small sample size and potentially non-normally distributed data.

Results

Regarding low-risk participants who received Defusion I: Relevant changes did not emerge and were not maintained at 4-month follow-up. The number of problematic behaviors decreased during and immediately after the protocol, but the number of problematic behaviors was not maintained at follow-up.

Regarding low-risk participants who received Defusion II: Participants showed statistically significant changes. Particularly, in acceptance of discomfort, they showed a low level of problematic behaviors even at follow-up (Pre-intervention: M = 2.75, SD = 2.22; Post-defusion: M = 0.40, SD = 0.89, Z = -1.63, p = .05, d = 1.26; 4-month FU: M = 0.00, SD = 0.00, Z = -1.84, p = .03, d = 1.60). All five low-risk participants who received Defusion II protocol achieved zero problematic behaviors at follow-up. Changes in psychological inflexibility (AFQ-S) were also significant (Pre: M = 5.00, SD = 2.21; Post-defusion: M = 3.05, SD = 1.07, Z = -2.03, p = .02, d = 1.12; 4-month FU: M = 1.20, SD = .60).

Regarding the six high-risk participants (Defusion II): Participants showed relevant reductions in number of problematic behaviors (Pre: M = 11.17, SD = 3.71; Post-defusion: M = 7.33, SD = 3.83, Z = -2.21, p = .01, d = 1.02; 4-month FU: M = 7.50, SD = 5.68, Z = -2.00, p = .02, d = 0.77). Although all but one participant showed reductions (one did not maintain the reduction at 4-month follow-up), these changes were statistically significant both at post-defusion and follow-up.

Likewise, accepting without judgment scores (KIMS) increased significantly at post-defusion (Post-values: M = 5.08, SD = 1.11; Post-defusion: M = 8.88, SD = 1.10, Z = -1.90, p = .02, d = 1.56), although the difference was attenuated at follow-up (M = 6.17).

No changes were observed in psychological inflexibility (AFQ-S) at post-defusion and follow-up (Pre: M = 4.88, SD = 2.14; Post-defusion: M = 4.71, SD = .86; 4-month FU: M = 4.96, SD = 1.09).

Discussion and conclusions

The pattern of results suggests that the effects of Defusion I protocol (with Type B interactions only) were limited in reducing problematic behaviors, which were not maintained over time, and with very small changes in psychological inflexibility (AFQ-S) and accepting without judgment (KIMS).

In contrast, Defusion II protocol showed reduction to zero level of problematic behaviors in all low-risk participants who received it only, and results were partly replicated in the effect obtained with high-risk participants. The limitations of the present study should be considered with caution. For example, the small number of participants prevented evaluation of each Defusion protocol with high-risk participants. School constraints forced protocol implementation by tutorial class, which precluded random distribution of participants to the two conditions.

However, this is the first occasion on which an RFT analysis of defusion in ACT has been conducted. Specifically, this preliminary analysis of defusion has differentiated deictic interactions (Type B) from deictic-plus-hierarchy (Type C) and regulation (Type D) interactions to alter the dominance of literal functions of thoughts and feelings. Future research is recommended to overcome the limitations of this study and extend and refine this RFT analysis of defusion methods.

Significance and contribution

This study constitutes the first RFT analysis of defusion interactions in ACT. It provides a preliminary but important analysis of how different types of relational interactions (deictic, hierarchical with examples, and regulatory) can differentially impact defusion outcomes. The findings suggest that defusion is not a singular process but involves multiple types of relational transformations of functions that must be appropriately sequenced and integrated for optimal effect.

The most important contribution is theoretical-methodological: it demonstrates how RFT provides tools for analyzing and refining ACT techniques in terms of underlying relational processes. This analytical approach has implications for improving the efficacy of defusion methods and potentially other ACT techniques analyzed analogously. The study opens a research line in which can be more precisely specified what types of interactions in defusion are critical for what types of behavioral or cognitive changes.