Acceptance and mindfulness as mediators of change in an acceptance-based behavior therapy and a non-directive supportive therapy for generalized anxiety disorderABSTRACT
Objective: Acceptance and mindfulness have been focused on as core clinical targets of evidence-based psychotherapeutic approaches for generalized anxiety disorder (GAD), such as acceptance-based behavior therapy (ABBT).
Methods: In the present study, we examined whether experiential avoidance (EA) and the five facets of mindfulness mediated clinical change in GAD core symptoms (i.e., worry and generalized anxiety) in 92 patients who received 10 sessions of either group ABBT or nondirective supportive group therapy (NDST). A mediation model was used to examine the relative magnitude of the indirect effects associated with each of the proposed mediators.
Results: ABBT, relative to NDST, had a significantly larger improvement on two mindfulness facets (describing and non-judgment) from pre- to mid-treatment. Several process variables were significantly associated with changes in anxiety and worry levels between midtreatment and posttreatment. Nevertheless, only one mindfulness facet (Describing) acted as a mediator for the later reduction in one of the outcomes (i.e., generalized anxiety).
Conclusions: Our findings are similar to those from previous research using time-lag analysis. There is a need to explore and test the relations between these and competing processes in relation to GAD symptoms in both controlled experiment settings and larger controlled trials.
KEYWORDS
Generalized Anxiety Disorder; Cognitive Behavioral Therapy; Mindfulness; Psychotherapy, Group; Mediation Analysis
SUMÁRIO
Objetivo: Aceitação e mindfulness têm sido focados como alvos clínicos centrais de abordagens psicoterapêuticas baseadas em evidências para o transtorno de ansiedade generalizada (TAG), como a terapia comportamental baseada em aceitação (TCBA).
Métodos: No presente estudo, examinamos se a evitação experiencial (EE) e as cinco facetas de mindfulness mediaram a mudança clínica nos sintomas centrais do TAG (i.e., preocupação e ansiedade generalizada) em 92 pacientes que receberam 10 sessões de TCBA em grupo ou de terapia de apoio não diretiva em grupo (TAND). Um modelo de mediação foi utilizado para examinar a magnitude relativa dos efeitos indiretos associados a cada um dos mediadores propostos.
Resultados: TCBA, em comparação a TAND, apresentou uma melhora significativamente maior em dois aspectos do mindfulness (descrever e não julgamento) do pré ao meio do tratamento. Diversas variáveis de processo foram significativamente associadas às mudanças nos níveis de ansiedade e preocupação entre o meio do tratamento e o pós-tratamento. Entretanto, apenas uma faceta de mindfulness (Descrever) atuou como mediadora da redução posterior em um dos desfechos (ansiedade generalizada).
Conclusões: Nossos achados são semelhantes aos de pesquisas anteriores que utilizaram análise time-lag. Há necessidade de explorar e testar as relações entre esses e outros processos concorrentes em relação aos sintomas de TAG, tanto em contextos experimentais controlados quanto em ensaios controlados com amostras maiores.
PALAVRAS-CHAVE
Transtorno de Ansiedade Generalizada; Terapia Cognitivo-Comportamental; Atenção Plena; Psicoterapia de Grupo; Análise de Mediação
INTRODUCTION
Several psychotherapeutic approaches have been demonstrated to be efficacious and effective treatments for generalized anxiety disorder (GAD) including cognitive-behavioral therapy (CBT),1 and acceptance-based behavioral therapy (ABBT).2-4 Although frequently used as a control intervention5, non-directive supportive therapy (NDST) presents results significantly equivalent to other first-choice interventions.1 Nevertheless, the specific processes responsible for the clinical changes is still relatively unknown.6,7 Thus, mediation analysis of the putative mechanisms of change within psychotherapeutic interventions for a specific clinical condition is needed for further technical refinement and improvement of the treatment's efficacy and effectiveness.8 This is particularly important in disorders with lower response rates to psychotherapy, such as GAD.1 Therefore, the overall goal of the present study is to evaluate potential mechanisms of change in psychotherapeutic approaches for individuals suffering from GAD.
The Acceptance-Based Model (ABM) and Acceptance-Based Behavior Therapy's (ABBT) processes of change
According to ABM, individuals with GAD develop a problematic way of relating to their own inner experiences (i.e, sensations, thoughts, emotions, and memories), resulting in increased efforts to change the way they feel.9 This pattern of struggling with one's own inner experiences is called experiential avoidance (EA).10,11 According to ABM, EA constitutes the main psychopathogenic factor in GAD,12 both in the production of distress and in functional impairment (i.e., behavioral restriction). Therefore, a central goal of ABBT, an ABM based psychotherapeutic treatment approach, is to decrease EA through the cultivation of opposite process of acceptance through increasing mindfulness.9
Acceptance is defined as the open and active willingness to stay in contact with one's own inner experiences, even if undesirable, allowing them to unfold without trying to control them.13 An often applied intervention to increase acceptance in ABBT is mindfulness meditation practice.14 Mindfulness can be broadly understood as an awareness that emerges from deliberate, nonjudgmental attention to the present moment.15 Mindfulness and acceptance practices are a central part of numerous behavior therapeutic developments (e.g., Dialectic Behavior Therapy;16 Acceptance and Commitment Therapy;13 and Meta Cognitive Therapy)17 targeting a range of psychiatric problem areas with promising results.3
It has been argued that the description or definition of mindfulness based on Kabat-Zinn15 does not contemplate the full complexity of the construct. Baer et al.18 developed an instrument widely adopted in clinical psychotherapy trials where they define mindfulness as a multifactorial construct composed of five different facets (i.e., Five Facet Mindfulness Questionnaire - FFMQ). The authors argued that identifying the facets that theoretically make up the construct of mindfulness would help to improve the understanding and identification of the relationships between mindfulness and other variables of interest (e.g., anxious symptoms). Thus, with mindfulness being a multidimensional construct, each dimension could be differentially related to specific clinical populations (e.g., people with GAD). Acceptance and mindfulness are considered core processes in the mechanism of clinical change of ABBT and have been demonstrated to mediate clinical outcomes in secondary analyses of data from previous ABBT trials targeting GAD.3,19
Mechanisms of change
ABBT is based on the acceptance model of GAD12. According to this model, individuals with GAD show a complex psychological pattern involving (a) an attention bias to threat; (b) a critical, judgmental reactivity toward their emotional responses and thoughts; and (c) worries as a way to distract from more distressing mental images and internal experiences.
Even though it is theoretically plausible that the five facets of mindfulness may contribute to improving GAD symptoms through ABBT (e.g., non-judging and non-reactivity promoting enhancement of (b) above), it is still unknown which of those processes are actually mediating clinical response.
Furthermore, although some researchers have suggested that there are specific psychological mechanisms yielding clinical improvement by each efficacious therapy in GAD,20-22 mediation analysis from some randomized controlled trials (RCTs) suggest otherwise. Research shows that despite targeting specific processes different efficacious approaches might be activating a common mechanism of change. For example, research has shown that both experiential avoidance/acceptance and decentering may lead to clinical change across intervention approaches.3,19,23,24 All these studies have compared acceptance-based approaches with other interventions, but none of them used NDST or any other non-CBT approach as a control, raising the issue of whether overlap across protocols could explain these findings.
As shown by the last Cochrane meta-analysis on psychotherapeutic treatments for GAD,5 patients have similar responses to different psychotherapeutic interventions. Thus, identifying potential common mechanisms between different approaches can help to elucidate the elements of effective therapy for GAD. Although demonstrating outcome mediations by isolated processes is insufficient to establish a potential mechanism, it is a necessary step for the identification of essential components of therapy8.
In the present study, we examined whether EA and the five facets of mindfulness mediated clinical change in GAD core symptoms (i.e., worry and generalized anxiety), as a secondary analysis of the data from an RCT comparing ABBT to a non-directive supportive therapy (NDST), both delivered in a group setting (see the efficacy article4 for more detailed information). We hypothesize that self-reported acceptance and mindfulness will increase significantly during treatment and will mediate clinical outcomes in both conditions. Considering that the efficacy study showed improvements in these processes for both groups,4 and that ABBT addresses these processes directly, demonstrating better clinical outcomes than NDST in this sample, we also hypothesized that the magnitude of change in the identified mediator processes (if any) would be significantly greater in the ABBT condition compared to NDST.
METHODS
This study involves a secondary data analysis from an RCT involving 10 sessions of ABBT or NDST.4 The RCT was registered on clinicaltrials.gov: 51363615.4.0000.0068/NCT03930095. The registration was made retrospectively due to a mistake in the registration process. The study was conducted under the Declaration of Helsinki and Ethical Guidelines for Clinical Studies and was approved by the research ethics committee of the Medical School of the University of São Paulo (USP), Brazil. Results from the larger RCT indicated that clients in both ABBT and NDST experienced a reduction in symptoms across treatment and follow-up and this decrease in symptoms was more rapid for clients in the ABBT condition based on the DASS-stress, PSWQ and HAMA. The pretreatment to posttreatment d's ranged from 0.80 to 1.49 for the ABBT condition and 0.41 to 1.19 for the NDST condition.4
Participants
Participants were the 92 individuals who completed either ABBT (n=46) or NDST (n=46) as part of the RCT (see efficacy article4 for the full methods, sociodemographic data, and outcome results). Participants were recruited from a list of patients with a potential diagnosis of GAD who were enrolled in the outpatient service of a university hospital in São Paulo, Brazil. Participants were included if they met eligibility including having a diagnosis of GAD based on the MINI for DSM-IV.25 Other inclusion criteria included being between the ages of 18-65 and being literate. Patients were excluded if they had bipolar disorder, psychosis, substance dependence, or showed moderate to high suicidal risk based on the MINI.
Outcome measures
Brazilian version of the Penn State Worry Questionnaire (PSWQ):26 The PSWQ is an index of excessive and uncontrollable worry and includes 16 items.27 Each item is rated on a five-point scale, ranging from 1 (not at all typical to me) to 5 (very typical of me). A higher PSWQ score corresponds to a higher level of worry. The PSWQ has been shown to display high internal reliability as well as divergent, convergent, and discriminant validity.27 In the current sample, the internal consistency of the PSWQ was α = .91.
Brazilian version of the Depression, Anxiety, and Stress Scale - Stress subscale (DASS-Stress):28 The DASS – Stress subscale assesses the state of persistent arousal and tension with 7 items.29 Each item is scored using a four-point severity/frequency scale from 0 (did not apply to me at all) to 3 (applies to me very much, or most of the time). A higher DASS – Stress subscale score indicates a higher level of stress. The subscale displays adequate internal consistency and evidence of several forms of validity, including convergent validity.29 In the current sample, the internal consistency of the DASS – Stress subscale was α = .80.
Brazilian version of the Hamilton Anxiety Scale (HAM-A):30 The HAM-A is an interview-administered measure of anxiety, including cognitive and somatic symptoms, with 14 items.31 Each item is scored on a scale from 0 (not present) to 4 (severe). A higher score indicates a higher level of anxiety symptoms. In the current sample, the internal consistency of the HAM-A was α = .84.
Mediator Measures
Brazilian version of the Acceptance and Action Questionnaire-II (AAQ-II 7):32 The AAQ-II assesses experiential avoidance and psychological inflexibility.33 The AAQ-II displays adequate internal consistency, satisfactory test-retest reliability, and several types of validity (e.g., concurrent and convergent validity). In the current sample, the internal consistency of the AAQ-II 7 was satisfactory at α = 0.88.
Brazilian version of the Five-factor Mindfulness Questionnaire (FFMQ):34 The five facets proposed as components of mindfulness in the FFMQ are: observing (i.e., noticing internal and external experiences), describing (i.e., being able to express experiences in words), acting with awareness (i.e., being attentive to the ongoing activity without being distracted), nonjudging of internal experience (i.e., interacting with internal content without evaluating or criticizing it), and nonreacting to internal experience (i.e., allowing internal content to flow without being too reactive to it).18 In the original study, the five factor structure of the FFMQ was supported. Each of the five subscales of the FMMQ had acceptable internal consistencies (ά = .75-.91). Additionally, there were significant correlations with conceptually related measures, such as psychological symptoms and thought suppression. In the current study, the internal consistency was acceptable for four of the five FFMQ subscales (α = .75-.90), but the non-reactivity subscale had a lower alpha coefficient (α = .68).
Procedure
Randomization was conducted by an independent individual using the method of randomly permuted blocks35 and stratified into two blocks based on psychotropic medication use. Participants were allocated to receive 10 sessions of either ABBT or NDST, both administered in groups of 11 to 12 members with a therapist and a co-therapist. The 10 sessions were completed within 14 weeks. The first six sessions were weekly and the last four, biweekly. The mediators (AAQ-II and the five subscales of the FFMQ) were assessed at pre-treatment, mid-treatment (week 6), and post-treatment via questionnaires. The outcome variables (PSWQ, DASS-Stress, and HAMA) were completed by the participants at pretreatment (week 0), mid-treatment (week 6), and post-treatment (week 12) via questionnaires.
Acceptance-based Behavioral Therapy (ABBT): clients received 10 2-hour sessions of group therapy and one follow-up session. The ABBT protocol was adapted from Roemer and Orsillo36 and Heatherington et al.37 All elements of the original treatment protocol were retained in our adaptation, which integrated two new strategies: written exposure38 and a functional assessment based on the collective Matrix from Polk and Schoendorff39 Written imaginal exposure was utilized as a means of practicing and enhancing willingness to experience anxiety (rather than extinction). Using the Matrix, participants identified their most important undesirable experiences and thoughts, and commonly used strategies, to eliminate or attenuate them (i.e., experiential avoidance). Then, they focused on their life goals and values (approach motivation) by identifying actions consistently aligned with them.
Non-directive Supportive Therapy (NDST): clients in this also received 10 2-hour sessions of group therapy and one follow-up session on the same time course as those in ABBT. The NDST protocol was based on the standards for brief supportive psychotherapy proposed by Markowitz.40 Our NDST manual offered generic and nondirective psychoeducation on GAD symptoms, nosology, treatments, and epidemiology. The therapist was not allowed to conduct any psychoeducation on theoretical models, techniques, psychological mechanisms, or emotion‐regulation strategies.
Therapists and adherence: therapists in both treatment conditions were clinical psychologists with equivalent backgrounds and specialization in CBT. The average time of clinical experience of the therapists was 9.0 (SD=5.0) years in the ABBT condition and 12.8 (SD=9.8) in NDST. Two sessions from each therapy group were randomly selected and assessed for adherence to the respective protocols. As described in the efficacy article4 therapists showed good overall adherence to the protocols.
Statistical analysis
Missing data were imputed using multiple imputation (MI) with maximum likelihood estimation in SPSS. MI is a technique that has been shown to produce less bias in mediation parameter estimates and with adequate power.41 In the efficacy RCT4 we examined the effect of the pattern of missingness (completers vs. noncompleters) on the rate of change by treatment condition for each of the primary outcome variables. For each primary outcome variable, there were nonsignificant effects for Completer Status × Time × Treatment Group (p's from .36 to .98). In the current study, MI was performed prior to the mediation analyses in two steps: (1) the X, M, and Y variables were included in the imputation model in 20 separate datasets, and (2) the mediation models were fitted within each imputed dataset. Subsequently, estimates were combined across imputations. Prior to MI, assumptions for regression analyses (i.e., linearity, normality of residuals, homoscedasticity, and independence of errors) were checked and met.
Mediation analyses were carried out in two steps using the PROCESS mediation macro for SPSS. First, univariate analyses were executed in which each mediator was used separately. Second, multiple mediation regression analysis was pre-planned for use42_44; however, as can be seen in the results section, multiple analysis was not used due to the lack of significant mediators in the first step using univariate analyses. Regression analysis consisted of determining the effect of the independent variable (group assignment) on the proposed mediators (the AAQ-II and the FFMQ subscales), and the effect of the proposed mediators on each of the outcomes. (See also Figure 1 for the study model). A bootstrapping method was employed to determine indirect effects43 with n = 5000 bootstrap re-samples. This method is recommended for mediation analyses with small to moderate sample sizes.42 Bootstrapping is a non-parametric procedure that produces an approximation of the sampling distribution of the indirect effects. This is attained through empirically generating a sample (with replacement of size n = 5000) from the full data set and calculating the indirect effects in all the re-samples. Bootstrapping confidence intervals are preferred over the Sobel test because often the sampling distribution of the indirect effect is non-normal. Path estimates are calculated using OLS regression. As illustrated in Figure 1, path a is the direct effect of the independent variable (group) on the mediators, path b represents the direct effect of the mediators on the dependent variables. The indirect effect (or the test of mediation) is shown in path a × b, and this estimates the effects of group on the dependent variables through the mediators, while controlling for pretreatment levels of the dependent variables. The direct effect shown in path c’ is the effect of group on the dependent variables. As is recommended by Hesser,45 a reversed model was also examined to compare the strength of associations where early changes in the outcomes predicted subsequent changes in the mediators in order to provide support for the direction of the relationship.
RESULTS
Mediation analyses
The descriptive statistics for the two groups (ABBT and NDST) on the six mediators and three outcomes at pre-, mid-, and post-treatment are shown in Table 1. As is displayed in Tables 2-4 under "Effect of IV on M (a)", ABBT, relative to the control group, had a significantly larger increase (improvement) on two FFMQ subscales (describing and non-judgment) from pre- to mid-treatment.
Descriptive statistics across the two groups for the six proposed mediators and three outcomes at pre-, mid-, and post-treatment
Bootstrapped Point Estimates and Bias-Corrected Confidence Intervals for the Indirect Effects of Treatment Condition on the PSWQ: Single mediator analyses
Bootstrapped Point Estimates and Bias-Corrected Confidence Intervals for the Indirect Effects of Treatment Condition on the DASS - Stress: Single mediator analyses
Bootstrapped Point Estimates and Bias-Corrected Confidence Intervals for the Indirect Effects of Treatment Condition on the HAM-A: Single mediator analyses
Indirect effects were first explored on the PSWQ. Early improvement (pre- to mid-treatment) in AAQ-II and one FFMQ subscale (acting with awareness) were significantly associated with reduction (improvement) on the PSWQ from mid- to post-treatment. However, early change in the six mediators (pre- to mid-treatment) did not significantly act as mediators for the later change on the PSWQ (mid- to post-treatment). Instead, all six statistical models showed only evidence of direct effects.
Indirect effects were then examined using the DASS – Stress as an outcome. Pre- to mid-treatment changes on the six mediators were not significantly associated with reduction (improvement) on the DASS – Stress from mid- to post-treatment. Early change in the six mediators (pre- to mid-treatment) did not significantly act as mediators for the later change on the DASS – Stress (mid- to post-treatment). The statistical models did not show evidence of indirect effects.
Indirect effects were then explored on the HAMA. Early improvement (pre- to mid-treatment) on two of the FFMQ subscales (describing and non-reactivity) were significantly associated with reduction (improvement) on the HAMA from mid to post-treatment. The only mediator to display a significant indirect effect was one FFMQ subscale (describing), in that early change in describing from pre- to mid-treatment did significantly act as a mediator for the later change on the HAMA (mid- to post-treatment). Except for the mediational analysis on one FFMQ subscale (non-judging), the other statistical models showed evidence of direct effects.
Note that multiple mediation regression analysis was not used regarding the three outcomes (PSWQ, DASS – Stress, and HAMA) due to the fact that only one mediator (describing) was identified in the univariate analyses above.
Reversed model
To examine whether treatment improvements might mediate changes in the variables used as mediators above, follow-up analyses were conducted. These analyses were primarily executed to investigate which model out of the following two appeared to be the most plausible one: (1) initial model – the proposed mediators (AAQ-II and FFMQ subscales) were used as mediators and the outcome measures were employed as dependent variables in the statistical analyses (these results are reported above under the heading "Mediation analyses") and (2) reversed model – the outcome measures (PSWQ, DASS – Stress and HAMA) were used as mediators (pre- to mid-treatment) and the mediators in the initial model were employed as dependent variables (mid- to post-treatment) in the statistical analyses. Otherwise, the analyses were identical to those in the above section (Mediation analysis). The findings for the PSWQ, DASS – Stress, and HAMA were identical in terms of indirect effects: the indirect effect (ab) of the PSWQ, DASS – Stress, and HAMA, in the reversed model was not significant for AAQ-II and the five FFMQ subscales.
DISCUSSION
In the present study, acceptance and mindfulness processes were explored as potential mechanisms of change in ABBT and NDST treatments for GAD. Compared to NDST, ABBT resulted in significantly larger increases in two mindfulness processes, describing and non-judgment, from pre- to mid-treatment (a path). Irrespective of the assigned group, early improvement in acceptance and three mindfulness processes were significantly associated with improvement on two of the outcomes from mid- to post-treatment (b path). More specifically, changes in EA and one specific mindfulness process, ‘acting with awareness’, contributed significantly to the outcome regarding worry. Also, early improvement in the mindfulness processes ‘describing’ and ‘non-reactivity’ was significantly associated with changes in anxiety levels between mid-treatment and post-treatment. Last and most importantly, the findings indicated that early improvement in one process – describing – acted as a mediator for later reduction on one of the outcomes (i.e., HAMA) (ab path). These findings give partial support for our main hypothesis, since several acceptance and mindfulness processes were significantly improved and for some processes more so in the ABBT group compared to the NDST group, and because change in one mindfulness process acted as a mediator for one of the outcomes.
Similar results were found in a secondary analysis of an RCT comparing ABBT with Applied Relaxation (RA)46 where the data showed that increased decentering (a key feature of mindfulness) was associated with decreases in anxiety and that changes in decentering preceded changes in anxiety within both groups; even though mindfulness enhancing techniques were not an explicit part of AR. In further analysis from the same RCT dataset, Eustis et. al19 found a reduction in EA to partially mediate outcomes on worry and quality of life in both groups. Another RCT, comparing CBT with Acceptance and Commitment Therapy (ACT: an acceptance/mindfulness-based approach) for mixed anxiety disorders, found defusion/decentering to mediate changes across treatment conditions.24
The main hypothesis for these findings was that different techniques from these approaches may carry common clinical elements that could promote acceptance and mindfulness. However, as discussed by Hayes-Skelton et al46 and more recently by Wang et al.,47 although finding similar results across treatments provides support for a common mechanism, most of the studies compared treatments that have several common CBT techniques, including homework, self-monitoring, and repeated practice, and it is possible that other effective treatments out of the CBT branch may not promote acceptance or mindfulness. For example, only mindfulness-based intervention showed changes in decentering when compared with antidepressant medication for people with depression.48 In fact, none of these studies used therapies without any technique tailored to target putative pathological processes according to some cognitive-behavioral model (e.g., NDST) as a control. Thus, a particular strength of this study is that we examined EA and mindfulness across treatments with different theoretical backgrounds and few overlapping features. Finally, an important aspect to be considered is that while the previous studies focused on North American populations, this study found similar results in a South American population, contributing to the generalizability of these findings.
Another perspective on the current findings comes from a recent systematic review and meta-analysis on the mediators of acceptance and mindfulness-based therapies for anxiety and depression.49 In the study, three mediators were examined: mindful attention, decentering, and acceptance. Based on data from 33 studies, the overall mediating effect of the three mediators was small to medium across treatment types, control types, type of mediation analysis, and outcomes. In subsequent analyses, type of mediation analysis was shown to moderate the mediated effect. In trials that used correlation-based analyses, mediated effects were demonstrated. However, in studies using time-lag analyses, where changes in mediators must precede changes in outcomes, significant mediated effects were not shown. In light of the recent review of literature, it is thus possible that the use of time-lag analyses in the current study prevented us from demonstrating more or stronger mediational effects. However, alternative analyses, using the correlation-based approach, would have an important disadvantage: such analyses miss the possibility to establish temporal precedence, which is a vital piece to provide evidence of causality. In relation to the current study, it is also worth noting that the combination of a small-to-moderate sample size and the rigidity of time-lag analyses might have reduced the statistical power to detect existing, but potentially small, mediation effects.
Altogether these data suggest that EA and mindfulness processes may play a significant role as mediators of improvement in GAD, regardless of the fact of being or not directly or intentionally targeted by the specific psychotherapy protocol applied. Reductions in EA and enhancement of mindfulness facets seem to be important components of the complex and still quite unknown mechanisms of change of different psychotherapeutic interventions. As we have hypothesized, it is possible that the level of effectiveness of different psychotherapies for GAD would be related to the extent to which each intervention addresses these processes. Indeed, our data showed that ABBT had a significantly larger increase (improvement), compared to NDST, on two FFMQ subscales (describing and non-judgment). However, only the describing facet was shown to mediate improvements and only in anxiety symptoms but not worry. These findings raise a question: why would the describing facet mediate anxiety symptoms while non-judgment would not, considering that, in the AB model, non-acceptance is an important part of the problematic relationship with internal experiences? Furthermore, such a relationship is regarded as a core pathological process in GAD, directly related to experiential avoidance, and specifically targeted by ABBT interventions. One could speculate that describing skills are necessary for the further non-judgment of anxiety and other internal experiences. Thus, effective non-judgment skills would require knowing what exactly is not to be judged, and the ability to describe internal experiences would promote this knowledge. In other words, the describing facet would moderate the non-judgment clinical effects. In addition, even more speculatively, it is possible that responders of both groups showed increased acceptance; however, in ABBT group, a more accurate describing skill could have led to a higher quality or precision of non-judgment. Other studies are needed to investigate the potential role of the relationship between these and other mindful facets in mediating clinical outcomes.
This study findings also point to the necessity of research addressing how and to what extent each of these non-CBT/non-acceptance-based approaches indirectly affects these processes, and whether other specific processes would be more importantly implicated than EA and mindfulness in mediating the clinical changes produced by such interventions.
In the case of the mediation of EA and mindfulness being supported, research searching for different, simpler, and more effective ways (techniques) to promote positive changes in these processes would contribute to the development of more effective psychotherapeutic interventions for GAD. For example, it is possible that by targeting some of the elements from approaches like NDST (e.g., space to speak freely about emotions, and therapist responsiveness to their expression) may also bring new ways of targeting these underlying processes and may potentially improve the effectiveness of current protocols.
Although the present study found experiential avoidance/acceptance and some mindfulness facets to partially mediate clinical outcome in two different psychotherapeutic treatments for GAD, we cannot say whether changes in these processes would predict outcomes above and beyond other hypothesized mechanisms. As mentioned before, no single study will identify all the processes potentially encompassing a mechanism. Other processes can be found to play a central role in the mechanisms of change in different treatments for anxiety (e.g., extinction, reduced physical tension, reappraisal, metacognition or intolerance of uncertainty).50 For example, two studies51,52 found that not only changes in acceptance of internal experiences but also engagement in valued action predicted responder status over and above changes in worry for clients receiving ABBT for GAD. Well-tailored research is needed to verify whether changes in EA/acceptance and mindfulness predict outcomes above and beyond these other hypothesized mechanisms and whether other CBT effective protocols for GAD, targeting different processes, share the same mediators.
To address these hypotheses, multiple lines are needed at different levels of investigation: from experiments testing these processes as independent variables in more controlled settings (i.e., laboratory) to large sample and wide scope RCTs using mediation analysis of these and other putative key process variables investigating several psychological treatments for GAD simultaneously. Also needed are experiments specifically designed to test the mediators on core GAD symptoms and psychometric studies on the measurement of these process variables to establish measures with acceptable reliability and validity.
Limitations
This study has several limitations. One limitation is the lack of a moderator analysis to test who would benefit from what treatment. This is particularly relevant given the meta-analysis on mediators of acceptance and mindfulness-based therapies for anxiety and depression where the type of mediation analysis was the only statistically significant moderator.49 Only studies using correlation-based mediation approaches showed statistically significant mediating effects, while studies using causal time-lag analyses like we did, did not. Also, the lack of a valued action measure is another important limitation of our study, as engagement in valued action is a central target of ABBT protocols playing a central putative role in clinical change. Another limitation of the current study's analyses is that significant findings may have emerged as chance findings, due to the relatively high number of tests. As a result, false-negatives (Type II-error) are an alternative interpretation of the analytical findings. Finally, since there were no published data at the start of the study to use in the estimation of an appropriate sample size, the current study was not powered to detect differential mediated effects of ABBT and NDST. As a result, the current findings need to be interpreted cautiously. As a result, conclusions should be drawn with caution regarding the role of the psychological processes examined in this study.
CONCLUSION
Overall, we found little support in this study for acceptance and mindfulness processes acting as mediators in the treatment of GAD. The present study showed that only one mindfulness facet (describing) was supported as a mediator. Based on these findings, there is a need to further explore and test the processes identified in the current study, as well as other theory-based processes, in relation to GAD symptoms in both more controlled experimental settings and larger controlled trials.
FUNDING
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon request.
ACKNOWLEDGEMENTS
We are profoundly grateful to our colleagues from the Anxiety Disorders Program, AMBAN, for their highly professional voluntary and committed support on data collection. We are also thankful to Dr. Lizabeth Roemer for her kind support, since the beginning of the whole research project.
REFERENCES
-
1 Papola D, Miguel C, Mazzaglia M, Franco P, Tedeschi F, Romero SA, Patel AR, Ostuzzi G, Gastaldon C, Karyotaki E, Harrer M, Purgato M, Sijbrandij M, Patel V, Furukawa TA, Cuijpers P, Barbui C. Psychotherapies for Generalized Anxiety Disorder in Adults: A Systematic Review and Network Meta-Analysis of Randomized Clinical Trials. JAMA Psychiatry. 2024 Mar 1;81(3):250-259. doi: 10.1001/jamapsychiatry.2023.3971. Erratum in: JAMA Psychiatry. 2024 Mar 1;81(3):320. doi: 10.1001/jamapsychiatry.2023.5480. PMID: 37851421; PMCID: PMC10585589.
» https://doi.org/10.1001/jamapsychiatry.2023.3971» https://doi.org/10.1001/jamapsychiatry.2023.5480 -
2 Roemer L, Orsillo SM, Salters-Pedneault K. Efficacy of an acceptance-based behavior therapy for generalized anxiety disorder: evaluation in a randomized controlled trial. J Consult Clin Psychol. 2008 Dec;76(6):1083-9. doi: 10.1037/a0012720. PMID: 19045976; PMCID: PMC2596727.
» https://doi.org/10.1037/a0012720 -
3 Hayes-Skelton SA, Roemer L, Orsillo SM. A randomized clinical trial comparing an acceptance-based behavior therapy to applied relaxation for generalized anxiety disorder. J Consult Clin Psychol. 2013 Oct;81(5):761-73. doi: 10.1037/a0032871. Epub 2013 May 6. PMID: 23647281; PMCID: PMC3783580.
» https://doi.org/10.1037/a0032871 -
4 Sampaio TPDA, Jorge RC, Martins DS, Gandarela LM, Hayes-Skelton S, Bernik MA, Lotufo-Neto F. Efficacy of an acceptance-based group behavioral therapy for generalized anxiety disorder. Depress Anxiety. 2020 Dec;37(12):1179-1193. doi: 10.1002/da.23021. Epub 2020 Apr 25. PMID: 32333486.
» https://doi.org/10.1002/da.23021 -
5 Hunot V, Churchill R, Silva de Lima M, Teixeira V. Psychological therapies for generalised anxiety disorder. Cochrane Database Syst Rev. 2007 Jan 24;2007(1):CD001848. doi: 10.1002/14651858.CD001848.pub4. PMID: 17253466; PMCID: PMC7025441.
» https://doi.org/10.1002/14651858.CD001848.pub4 -
6 Goldfried MR. What should we expect from psychotherapy? Clin Psychol Rev. 2013 Apr 8;33(5):654-662. doi: 10.1016/j.cpr.2012.09.006. Epub ahead of print. PMID: 23628909.
» https://doi.org/10.1016/j.cpr.2012.09.006 -
7 Levin ME, Luoma JB, Haeger JA. Decoupling as a mechanism of change in mindfulness and acceptance: a literature review. Behav Modif. 2015 Nov;39(6):870-911. doi: 10.1177/0145445515603707. Epub 2015 Sep 8. PMID: 26349756.
» https://doi.org/10.1177/0145445515603707 -
8 Kazdin AE. Mediators and mechanisms of change in psychotherapy research. Annu Rev Clin Psychol. 2007;3:1-27. doi: 10.1146/annurev.clinpsy.3.022806.091432. PMID: 17716046.
» https://doi.org/10.1146/annurev.clinpsy.3.022806.091432 -
9 Roemer L, Orsillo SM. Expanding Our Conceptualization of and Treatment for Generalized Anxiety Disorder: Integrating Mindfulness/Acceptance-Based Approaches With Existing Cognitive-Behavioral Models. Clin Psychol [Internet]. 11 maio 2006 [citado 3 out 2024];9(1):54-68. Disponível em: https://doi.org/10.1093/clipsy.9.1.54
» https://doi.org/10.1093/clipsy.9.1.54 -
10 Hayes SC, Wilson KG, Gifford EV, Follette VM, Strosahl K. Experimental avoidance and behavioral disorders: a functional dimensional approach to diagnosis and treatment. J Consult Clin Psychol. 1996 Dec;64(6):1152-68. doi: 10.1037//0022-006x.64.6.1152. PMID: 8991302.
» https://doi.org/10.1037//0022-006x.64.6.1152 -
11 Berman NC, Wheaton MG, McGrath P, Abramowitz JS. Predicting anxiety: the role of experiential avoidance and anxiety sensitivity. J Anxiety Disord. 2010 Jan;24(1):109-13. doi: 10.1016/j.janxdis.2009.09.005. PMID: 19819106.
» https://doi.org/10.1016/j.janxdis.2009.09.005 -
12 Roemer L, Salters K, Raffa SD, Orsillo SM. Fear and avoidance of internal experiences in GAD: Preliminary tests of a conceptual model. Cogn Ther Res. 2005;29:71-88. doi: 10.1007/s10608-005-1650-2.
» https://doi.org/10.1007/s10608-005-1650-2 - 13 Hayes SC, Strosahl K, Wilson KG. Acceptance and commitment therapy: an experiential approach to behavior change. New York: Guilford Press; 1999. 304 p.
- 14 Roemer L, Orsillo SM. Mindfulness and acceptance-based behavioral therapies in practice. New York: Guilford Press; 2009.
- 15 Kabat-Zinn J. Full catastrophe living: using the wisdom of your body and mind to face stress, pain, and illness. New York: Delta; 1990.
- 16 Linehan MM. Cognitive-behavioral treatment of borderline personality disorder. New York: Guilford Press; 1993.
-
17 Wells A. Meta-cognition and worry: a cognitive model of generalized anxiety disorder. Behav Cogn Psychother. 1995;23:301-20. doi: 10.1017/S1352465800015897.
» https://doi.org/10.1017/S1352465800015897 - 18 Baer RA, Smith GT, Hopkins J, Krietemeyer J, Toney L. Using self-report assessment methods to explore facets of mindfulness. Assessment. 2006;13(1):27-45.
-
19 Eustis EH, Hayes-Skelton SA, Roemer L, Orsillo SM. Reductions in experiential avoidance as a mediator of change in symptom outcome and quality of life in acceptance-based behavior therapy and applied relaxation for generalized anxiety disorder. Behav Res Ther. 2016;87:188-95. doi: 10.1016/j.brat.2016.09.012.
» https://doi.org/10.1016/j.brat.2016.09.012 -
20 Forman EM, Herbert JD, Moitra E, Yeomans PD, Geller PA. A randomized controlled effectiveness trial of acceptance and commitment therapy and cognitive therapy for anxiety and depression. Behav Modif. 2007;31(6):772-99. doi: 10.1177/0145445507302202.
» https://doi.org/10.1177/0145445507302202 -
21 Forman EM, Shaw JA, Goetter EM, Herbert JD, Park JA, Yuen EK. Long-term follow-up of a randomized controlled trial comparing acceptance and commitment therapy and standard cognitive behavior therapy for anxiety and depression. Behav Ther. 2012;43(4):801-11. doi: 10.1016/j.beth.2012.04.004.
» https://doi.org/10.1016/j.beth.2012.04.004 -
22 Donegan E, Dugas MJ. Generalized anxiety disorder: a comparison of symptom change in adults receiving cognitive-behavioral therapy or applied relaxation. J Consult Clin Psychol. 2012;80(3):490-6. doi: 10.1037/a0028132.
» https://doi.org/10.1037/a0028132 -
23 Arch JJ, Craske MG. Acceptance and commitment therapy and cognitive behavioral therapy for anxiety disorders: Different treatments, similar mechanisms? Clin Psychol Sci Pract. 2008;15(4):263-79. doi: 10.1111/j.1468-2850.2008.00137.x.
» https://doi.org/10.1111/j.1468-2850.2008.00137.x -
24 Arch JJ, Wolitzky-Taylor KB, Eifert GH, Craske MG. Longitudinal treatment mediation of traditional cognitive behavioral therapy and acceptance and commitment therapy for anxiety disorders. Behav Res Ther. 2012;50(7-8):469-78. doi: 10.1016/j.brat.2012.04.007.
» https://doi.org/10.1016/j.brat.2012.04.007 -
25 Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998;59 Suppl 20:22-33. Available from: http://www.ncbi.nlm.nih.gov/pubmed/9881538
» http://www.ncbi.nlm.nih.gov/pubmed/9881538 -
26 Castillo C, Macrini L, Cheniaux E, Landeira-Fernandez J. Psychometric properties and latent structure of the Portuguese version of the Penn State Worry Questionnaire. Span J Psychol. 2010;13(1):431-43. doi: 10.1017/s113874160000398x.
» https://doi.org/10.1017/s113874160000398x - 27 Meyer TJ, Miller ML, Metzger RL, Borkovec TD. Development and validation of the Penn State Worry Questionnaire. Behav Res Ther. 1990;28(6):487-95.
-
28 Vignola RC, Tucci AM. Adaptation and validation of the depression, anxiety and stress scale (DASS) to Brazilian Portuguese. J Affect Disord. 2014;155:104-9. doi: 10.1016/j.jad.2013.10.031.
» https://doi.org/10.1016/j.jad.2013.10.031 - 29 Lovibond PF, Lovibond SH. The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behav Res Ther. 1995;33(3):335-43.
-
30 Kummer A, Cardoso F, Teixeira AL. Generalized anxiety disorder and the Hamilton Anxiety Rating Scale in Parkinson's disease. Arq Neuropsiquiatr. 2010;68(4):495-501. doi: 10.1590/S0004-282X2010000400005.
» https://doi.org/10.1590/S0004-282X2010000400005 - 31 Hamilton M. The assessment of anxiety states by rating. Br J Med Psychol. 1959;32(1):50-5.
-
32 Barbosa LM, Murta SG. Propriedades psicométricas iniciais do Acceptance and Action Questionnaire - II - versão brasileira. Psico-USF. 2015;20(1):75-85. doi: 10.1590/1413-82712015200107.
» https://doi.org/10.1590/1413-82712015200107 -
33 Bond FW, Hayes SC, Baer RA, Carpenter KM, Guenole N, Orcutt HK, Waltz T, Zettle RD. Preliminary psychometric properties of the Acceptance and Action Questionnaire-II: a revised measure of psychological inflexibility and experiential avoidance. Behav Ther. 2011;42(4):676-88. doi: 10.1016/j.beth.2011.03.007.
» https://doi.org/10.1016/j.beth.2011.03.007 -
34 Barros VV, Kozasa EH, Souza ICW, Ronzani TM. Evidências de validade da versão brasileira do Questionário das Cinco Facetas de Mindfulness (FFMQ-BR). Psicol Teor Pesq. 2014;30(3):317-27. doi: 10.1590/S0102-37722014000300009.
» https://doi.org/10.1590/S0102-37722014000300009 -
35 Matts JP, Lachin JM. Properties of permuted-block randomization in clinical trials. Controlled Clinical Trials. 1988;9(4):327–344. doi:10.1016/0197-2456(88)90046-3.
» https://doi.org/10.1016/0197-2456(88)90046-3 -
36 Roemer L, Orsillo SM. An open trial of an acceptance-based behavior therapy for generalized anxiety disorder. Behav Ther. 2007;38(1):72-85. doi: 10.1016/j.beth.2006.04.004.
» https://doi.org/10.1016/j.beth.2006.04.004 -
37 Heatherington L. Applying Group Cognitive Behavioral Therapy for Anxiety Disorders in Community Settings: Retention, Outcome, and Clinical Considerations. J Cogn Psychother. 2014;28(2):117-33. doi: 10.1891/JCP.28.2.
» https://doi.org/10.1891/JCP.28.2 - 38 Dugas MJ, Robichaud M. Cognitive-behavioral treatment for generalized anxiety disorder: from science to practice. New York: Routledge; 2007.
- 39 Polk KL, Schoendorff B, editors. The ACT matrix: A new approach to building psychological flexibility across settings and populations. Oakland, CA: Context Press/New Harbinger Publications; 2014.
-
40 Markowitz JC. What is Supportive Psychotherapy? Focus (Am Psychiatr Publ). 2014;12(3):285-9. doi: 10.1176/appi.focus.12.3.285.
» https://doi.org/10.1176/appi.focus.12.3.285 - 41 Zhang Z, Wang L. Methods for mediation analysis with missing data. Psychometrika. 2013;78(1):154-84.
- 42 Hayes AF. Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Commun Monogr. 2009;76(4):408-20.
- 43 Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods. 2008;40(3):879-91.
- 44 Rucker DD, Preacher KJ, Tormala ZL, Petty RE. Mediation analysis in social psychology: Current practices and new recommendations. Soc Personal Psychol Compass. 2011;5(6):359-71.
- 45 Hesser H. Modeling individual differences in randomized experiments using growth models: recommendations for design, statistical analysis, and reporting of results of internet interventions. Internet Interv. 2015;2(2):110-120.
-
46 Hayes-Skelton SA, Calloway A, Roemer L, Orsillo SM. Decentering as a potential common mechanism across two therapies for generalized anxiety disorder. J Consult Clin Psychol. 2015;83(2):395-404. doi: 10.1037/a0038305.
» https://doi.org/10.1037/a0038305 -
47 Wang Z, Tang X, Hu A, Chiu W, Hofmann SG, Liu X. Decentering as a mediator of the effect of mindfulness on emotional distress: Evidence from cross-sectional and longitudinal designs. Psychother Res. 2025 Nov;35(8):1458-1472. doi: 10.1080/10503307.2024.2426562. Epub 2024 Nov 17. PMID: 39550767.
» https://doi.org/10.1080/10503307.2024.2426562 -
48 Bieling PJ, Hawley LL, Bloch RT, Corcoran KM, Levitan RD, Young LT, et al. Treatment-specific changes in decentering following mindfulness-based cognitive therapy versus antidepressant medication or placebo for prevention of depressive relapse. J Consult Clin Psychol. 2012;80(3):365-72. doi: 10.1037/a0027483.
» https://doi.org/10.1037/a0027483 -
49 Johannsen M, Nissen ER, Lundorff M, O’Toole MS. Mediators of acceptance and mindfulness-based therapies for anxiety and depression: A systematic review and meta-analysis. Clin Psychol Rev. 2022;94:102156. doi: 10.1016/j.cpr.2022.102156.
» https://doi.org/10.1016/j.cpr.2022.102156 -
50 Costa MA, Russell TA, Gosmann NP, Gonçalves F, Tatton-Ramos T, de Oliveira FB, Manfro GG. Mechanisms of improvement in generalized anxiety disorder: A mediation and moderated mediation analysis from a randomized controlled trial. Br J Clin Psychol. 2023 Mar;62(1):196-208. doi: 10.1111/bjc.12402. Epub 2022 Nov 29. PMID: 36447332.
» https://doi.org/10.1111/bjc.12402 - 51 Hayes SA, Orsillo SM, Roemer L. Changes in proposed mechanisms of action in an acceptance-based behavior therapy for generalized anxiety disorder. Behav Res Ther. 2010;48(3):238-45.
-
52 Marando-Blanck S, Hayes-Skelton SA, Roemer L, Orsillo SM. Examining interrelations among trajectories of mindful awareness, acceptance, and values-consistent actions in acceptance-based behavioral therapy for generalized anxiety disorder. Cogn Behav Ther. 2025 Sep;54(5):577-595. doi: 10.1080/16506073.2024.2423654. Epub 2024 Nov 7. PMID: 39508504; PMCID: PMC12056163.
» https://doi.org/10.1080/16506073.2024.2423654
