Accepted Manuscript Title: Mediators of Transdiagnostic Group Cognitive Behavior Therapy Author: Alexander M. Talkovsky Peter J. Norton PII: DOI: Reference:
S0887-6185(14)00141-8 http://dx.doi.org/doi:10.1016/j.janxdis.2014.09.017 ANXDIS 1652
To appear in:
Journal of Anxiety Disorders
Received date: Revised date: Accepted date:
30-4-2014 15-9-2014 26-9-2014
Please cite this article as: Talkovsky, A. M., and Norton, P. J.,Mediators of Transdiagnostic Group Cognitive Behavior Therapy, Journal of Anxiety Disorders (2014), http://dx.doi.org/10.1016/j.janxdis.2014.09.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Peter J. Norton, Ph.D. 1 University of Houston
Ac ce p
te
d
M
an
us
1
cr
Alexander M. Talkovsky, M.A. 1
ip t
Mediators of Transdiagnostic Group Cognitive Behavior Therapy
Page 1 of 29
RUNNING HEAD: Mediators of Transdiagnostic CBT
1
Journal of Anxiety Disorders
an
us
September 15, 2014
cr
Submitted to:
ip t
Mediators of Transdiagnostic Group Cognitive Behavior Therapy
Word Count: 5804
Ac ce p
te
d
M
Tables and Figures: 1
Page 2 of 29
Mediators of Transdiagnostic CBT
2 Abstract
The efficacy of cognitive-behavioral therapy (CBT) for anxiety is well established. Investigations into the mechanisms of change in CBT report changes in cognitive vulnerabilities
ip t
mediating improvements over the course of treatment. As anxiety disorders share certain risk factors, there is a trend towards CBT emphasizing these vulnerabilities, including negative
cr
affectivity (NA) and also more specific constructs such as anxiety sensitivity (AS) and
us
intolerance of uncertainty (IU). The purpose of this investigation was to analyze potential mediators of anxiety reduction over the course of transdiagnostic group CBT. NA, AS, and IU all
an
decreased over the course of treatment. Among the potential mediators, change in NA had a significant relationship with change in anxiety but change in AS and change in IU did not.
M
Neither the main effect of primary diagnosis nor the interactions between potential mediators and
d
primary diagnoses were significant, indicating that there were no differential changes in anxiety
te
or the potential mediators across primary diagnoses. Results strongly point towards NA as an
Ac ce p
overarching mediator of anxiety reduction during transdiagnostic group CBT.
Keywords: anxiety, transdiagnostic, cognitive behavior therapy, mediation
Page 3 of 29
Mediators of Transdiagnostic CBT
3
Mediators of Transdiagnostic Group Cognitive Behavior Therapy The most common class of psychological disorders is anxiety, as roughly a quarter to a third of all people will meet criteria for an anxiety disorder in their lifetimes (Kessler et al., 2005;
ip t
Kessler et al., 2012). A large body of research supports the use of cognitive behavioral therapy (CBT) for the treatment of anxiety (Norton & Price, 2007). CBT-based interventions have robust
cr
effects on panic (Barlow, Gorman, Shear, & Woods, 2000), social phobia (Heimberg, 2002)
us
generalized anxiety disorder (GAD; Covin, Ouimet, Seeds, & Dozois, 2008), posttraumatic stress disorder (PTSD; Powers et al., 2010), and obsessive compulsive disorder (OCD; Whittal,
an
Thordarson, & McLean, 2005), and clinically significant gains are maintained long-term. Despite the large body of empirical support for CBT, practitioners often struggle to
M
deliver care to all those that would benefit from it (Shafran et al., 2009), partially because it is
d
time-consuming, costly, and inefficient to disseminate multiple disorder-specific manualized
te
CBT protocols (Clark, 2009). There is a trend towards anxiety treatment that emphasizes vulnerability shared across the anxiety disorders, as there is a large amount of overlap (Barlow,
Ac ce p
2000; Norton, 2006). In Clark and Watson’s (1991) tripartite model of anxiety and depression, depression and anxiety share the underlying vulnerability of negative affectivity, or a dispositional penchant for responding to negative stimuli with feelings of fear, anxiety, depression, guilt and self-dissatisfaction (Clark, Watson, & Mineka, 2004). Negative affectivity serves as a higher-order factor subsuming lower-order risk factors that confer more specific risk for certain disorders. Brown, Chorpita, and Barlow (1998) observed a model in which negative affectivity had a direct causal influence on GAD, panic disorder, OCD, and social phobia and this model had an excellent fit to their data. Watson (2005) suggested that a model of anxiety disorders should not only include factors that cut across diagnoses (e.g., negative affectivity), but
Page 4 of 29
Mediators of Transdiagnostic CBT
4
also those that are more specific to particular conditions. The most heavily studied constructs proposed to be specific factors may be anxiety sensitivity and intolerance of uncertainty. Anxiety sensitivity, or the fear of fearful and/or anxious symptoms due to the belief that
ip t
they may have adverse consequences (Reiss & McNally, 1985) is heavily implicated in panic disorder (McNally, 2002) and PTSD (Taylor, 1993), but scores are also elevated among
cr
individuals with GAD, social phobia, and OCD compared to non-anxious controls (Taylor,
us
Koch, & McNally, 1992). Furthermore, Hazen, Walker, and Eldridge (1996) found that anxiety sensitivity (AS) is responsive to anxiety treatment, and this improvement is related to the
an
reduction of anxiety symptoms. In motor vehicle accident victims, Fedoroff, Taylor, Asmundson, and Koch (2000) found, using regression analyses, that AS and pain severity were significant
M
predictors of both trauma symptoms prior to treatment and symptom reduction following
d
treatment. Smits, Powers, Cho, and Telch (2004) used the Baron and Kenny (1986) method to
te
test whether a reduction in fear of fear underlies improvement resulting from CBT for panic disorder. They observed full mediation of the relationship between the effects of CBT and
Ac ce p
changes in global impairment, and partial mediation on agoraphobia, anxiety, and panic frequency. Hofmann and colleagues (2007) substantiated these results, finding that panic-related catastrophic cognitions mediated treatment change in CBT. Few studies analyze the role of anxiety sensitivity in treatment of other anxiety disorders, but rather they focus on panic disorder (Smits, Powers, Cho, & Telch, 2004) or PTSD (Fedoroff, Taylor, Asmundson, & Koch, 2000), leaving a need to analyze the rest of the anxiety disorders. Boswell and colleagues (2013) observed elevated AS in patients with primary diagnoses of panic disorder, social phobia, GAD, and OCD, and all groups demonstrated AS improvements following interoceptive exposures, and lower levels of AS were associated with lower levels of
Page 5 of 29
Mediators of Transdiagnostic CBT
5
clinical severity in analyses collapsed across principal diagnoses. Thus, AS may represent a specific vulnerability to panic, but it may confer risk to the other anxiety disorders as well. Similarly, the construct of intolerance of uncertainty (IU) has been operationalized in
ip t
many different ways (Starcevic & Berle, 2006). For the purposes of this investigation, it will be defined as a characteristic tendency to find uncertainty regarding outcomes of an event aversive
cr
(Freeston, Rhéaume, Letarte, Dugas, & Ladouceur, 1994). Treatment specifically targeting IU in
us
individuals with GAD leads to clinically significant change immediately following treatment and gains are maintained at twelve-month follow up whether treating individuals (Ladouceur et al.,
an
2000) or groups (Dugas et al., 2003). Although it was originally hypothesized that IU is specific to GAD (Dugas, Gagnon, Ladouceur, & Freeston, 1998), Carleton and colleagues (2012) found
M
elevated scores in social anxiety disorder (SAD), panic disorder (PD), GAD, OCD, and
d
depression patients, and concluded that IU may represent a transdiagnostic vulnerability.
te
McEvoy and Mahoney (2012) found that IU mediates the relationship between neuroticism, a higher order vulnerability, and social anxiety, panic, and depression; also, they found that the
Ac ce p
prospective anxiety factor of IU mediated the relationship between neuroticism and worry, and obsessive-compulsive symptoms. Belloch and colleagues (2011) found that IU decreased over the course of cognitive therapy for OCD. Mahoney and McEvoy (2012) observed reductions in IU specific to social situations following group CBT for SAD, but it is possible that this situation-specific assessment captures variance overlapping with diagnostic criteria for SAD and it would be worth replicating results with a more general assessment tool. IU is clearly relevant to the treatment process, but presently lacks empirical support as a mediator of treatment outcomes for anxiety. Despite its associations with GAD, IU is a promising transdiagnostic marker for anxiety in light of its associations with a number of other disorders.
Page 6 of 29
Mediators of Transdiagnostic CBT
6
Sexton, Norton, Walker, and Norton (2003) tested a hierarchical model in a sample of undergraduates with neuroticism serving as the first-order factor subsuming AS and IU. They found that AS acted as a risk for panic, obsessive-compulsive symptoms, and health anxiety, IU
ip t
acted as a risk for worry, and neuroticism was an overarching risk for all symptom types. Norton, Sexton, Walker, and Norton (2005) partially replicated these results in a sample of treatment-
cr
seekers. They observed that negative affectivity served as a global risk factor and influenced AS
us
and IU, AS affected panic, and health anxiety, and IU influenced worry and depression; however, the expected relationship between negative affectivity and obsessive-compulsive
an
symptoms was not observed. Norton and Mehta (2007) used structural equation modeling to compare potential hierarchical models. They improved the fit of previous models by including a
M
path from IU to obsessive-compulsive and social anxiety symptoms. These studies combine to
d
provide support for hierarchical models of anxiety with AS and IU acting as second-order risk
te
factors that show a degree of specificity, but not as much as some originally thought. As IU and AS are oft-acknowledged risk factors, and they decrease across treatment, it
Ac ce p
would be enlightening to determine their relationships to symptom reduction and to do so across diagnoses. Because of their demonstrated relationships with anxiety disorders and their response to treatment, this study posits AS and IU as potential mediators of anxiety reduction. Mediator variables explain the way an antecedent exerts its effects on a dependent variable, thus allowing a more complete understanding of how the criterion and outcome variables relate to one another (MacKinnon & Fairchild, 2009). Mediational analyses are important for treatment and prevention, as interventions are often designed to change an outcome by targeting mediators hypothesized to have a causal relationship with the outcome (MacKinnon, Fairchild, & Fritz, 2007). It is important to understand what makes therapy work because understanding
Page 7 of 29
Mediators of Transdiagnostic CBT
7
mechanisms of change can improve the identification of appropriate, better, different, and new treatments (Kazdin, 2007, 2009). Much of the extant literature suggests that cognitive variables mediate the outcomes of diagnosis-specific anxiety treatment (Arch, Wolitzky-Taylor, Eifert, &
ip t
Craske, 2012; Hofmann, 2004; Hofmann et al., 2007; Mahoney & McEvoy, 2012; Smits, Powers, Cho, & Telch, 2004).
cr
As interest in transdiagnostic psychotherapy grows, it is important to establish the role of
us
shared vulnerabilities, such as AS and IU, in the course of psychotherapy. In a mixed anxiety disorder sample of 35 individuals randomized to receive CBT that included components shared
an
across the anxiety disorders, Arch, Wolitzky-Taylor, Eifert, and Craske (2012) found that AS significantly mediated post-treatment worry and they observed no differential effects by
M
diagnosis. However, they failed to observe significant mediation for four of their five outcome
d
measures. Updating these results with a larger sample would contribute to the growing body of
te
transdiagnostic literature. We attempt to improve upon cross-sectional analysis by investigating simultaneous change over the entire course of a manualized treatment protocol. This study aims
Ac ce p
to identify potential mediators by observing the change in potential criterion, mediator, and outcome variables.
The purpose of this study is to analyze potential mediators of anxious symptom reduction in transdiagnostic cognitive behavioral group therapy. Norton (2008) established the effectiveness of transdiagnostic CBGT, but the mechanisms of change have not yet been established. This study seeks to identify potential mediators of overall anxiety reduction across anxiety diagnoses. The first hypothesis is that reductions in negative affectivity, a higher-order risk for anxiety, will be a significant mediator of symptom improvement across treatment in all diagnostic categories. Second, it is hypothesized that AS and IU will each act as specific
Page 8 of 29
Mediators of Transdiagnostic CBT
8
mediators, with AS significantly mediating improvement in panic symptoms and social anxiety, and IU mediating the improvement in worry. Methods
ip t
The data used in this study was collected during three previous clinical trials (Norton, 2008, 2012a; Norton & Barrera, 2012) that share methodologies and treatment protocols
cr
(Norton, 2012b). These results have not been previously published. All procedures were
us
approved for human subjects research by the Institutional Review Board at the University of Houston.
an
Participants
Participants included 256 individuals with principal diagnoses of panic disorder (PD),
M
social phobia (SP) or generalized anxiety disorder (GAD) seeking treatment at the Anxiety
d
Disorder Clinic (ADC) at the University of Houston. All of the individuals included in these
te
analyses were randomized into transdiagnostic cognitive behavioral group therapy. Over half (52.2%) of the sample identified as female. The mean age was 32.89 years (S.D. = 10.72). The
Ac ce p
sample was racially diverse; 56.1% self-identified as Caucasian, 9.6% as African-American, 7.4% as Asian or Asian-American, 20.7% as Hispanic or Latino(a), and 6.3% self-identified as multiracial or as a member as “Other.” The majority of the sample (57.7%) reported having completed at least a bachelor’s degree, 36.5% reported that they are married or cohabitating, and 76.2% reported that they are employed full time or are enrolled as full-time students. Income data were not collected. In this sample, 131 individuals (51.1%) had a primary diagnosis of SP. Among them, 31 (23.7%) had comorbid GAD, and 8 had comorbid PD (6.1%). Seventy eight individuals (30.5%) had a primary diagnosis of PD, and among them, 13 (16.7%) had comorbid SP and 24 (30.8%)
Page 9 of 29
Mediators of Transdiagnostic CBT
9
had comorbid GAD. Forty seven (18.4%) individuals had a primary diagnosis of GAD; 12 (25.5%) of them also had a diagnosis of PD, and 21 (44.7%) had a comorbid SP diagnosis. Comorbid depressive diagnoses were relatively common, but did not differ in frequency across
ip t
individuals with principal diagnoses of PD (32.1%), SP (27.5%), or GAD (44.7%), χ2 (2, n = 256) = 4.70, p = .10.
cr
Measures
us
Anxiety Disorders Interview Schedule for DSM-IV (ADIS-IV). The ADIS-IV (Brown, Di Nardo, & Barlow, 1994) is a semi-structured diagnostic interview that assess for the presence
an
and severity of anxiety, mood, somatoform; it also provides a brief screen for psychotic symptoms and substance abuse and dependence. Previous research on this sample reported high
M
interrater reliability and diagnostic agreement (86% agreement, κ = 0.759, p < .001; Chamberlain
d
& Norton, 2013).
te
State-Trait Anxiety Inventory – State Version (STAI-S). The STAI (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1993) is a 20-item self-report measure designed to assess
Ac ce p
state and trait levels of anxiety. Only the state version was used for the purposes of this investigation. Items are scored from 1 (“not at all”) to 4 (“very much so”) to indicate how much each statement reflects the subject’s present feelings; total scores range from 20 to 80. The original manual reported strong psychometric properties in several populations. Fisher and Durham (1999) reported that it is sensitive to treatment effects. In this sample, its internal consistency was excellent at session 1 (Cronbach α = .929). Positive and Negative Affect Schedule (PANAS). The PANAS (Watson, Clark, & Tellegen, 1988) is a 20-item self-report measure designed to assess positive affect (PA) and negative affect (NA). PA and NA are considered orthogonal measures of independent affective
Page 10 of 29
Mediators of Transdiagnostic CBT
10
dimensions (Watson, Clark, & Carey, 1988), but weak correlations are sometimes observed (Vittengl & Holt, 1998). This measure can assess either state or trait affect depending on the specific phrasing of the instructions (Watson, Clark, & Tellegen, 1988). For this study, the
ip t
instructions were phrased to measure trait affectivity. The PANAS has demonstrated acceptable psychometric properties in clinically anxious and depressed samples (Brown et al., 1998;
cr
Watson, Clark, & Carey, 1988), and good reliability in the current sample at pre-treatment
us
(Cronbach α = .821).
Anxiety Sensitivity Index (ASI). The ASI (Reiss, Peterson, Gursky, & McNally, 1986;
an
Peterson & Reiss, 1993) is a 16-item self-report measure designed to assess the fear of anxietyrelated symptoms. Items are scored from 0 (very little) to 4 (very much). Individuals with
M
anxiety disorder diagnoses score significantly higher than non-clinical controls, although the ASI
d
shows greater specificity for panic (Taylor, Koch, & McNally, 1992). It demonstrates adequate
te
criterion and construct validities (Reiss et al., 1986; Peterson & Reiss, 1993). The ASI demonstrated good internal consistency (Cronbach α = .865) at pre-treatment in this sample.
Ac ce p
Intolerance of Uncertainty Scale (IUS). The IUS (Freeston, Rhéaume, Letarte, Dugas, & Ladouceur, 1994) is a 27-item self-report measure designed to assess the extent to which an individual finds uncertainty to be aversive. It demonstrates adequate convergent and discriminant validity (Buhr & Dugas, 2002). The IUS demonstrated excellent internal consistency at pretreatment (Cronbach α = .957) in this sample. Procedures Prior to assessment, subjects provided their informed consent to participate in research and to receive treatment through the ADC. Then, they completed a large battery of self-report assessments, including the STAI, PANAS, ASI, and IUS as well as other measures not reltead to
Page 11 of 29
Mediators of Transdiagnostic CBT
11
this study. Next, subjects completed the ADIS-IV as administered by a trained graduate student and were subsequently provided a copy of their assessment report. Individuals with anxiety disorders that did not demonstrate dementia or other serious cognitive decline, adequately spoke
ip t
and read English, and did not display serious suicidality, substance use, or other conditions
requiring immediate intervention were invited to participate in a transdiagnostic group CBT
cr
program. The transdiagnostic model focuses on a global formulation of dysfunctional anxiety
us
rather than diagnosis-specific conceptualization.
The treatment program consisted of 12 weekly 2-hour long sessions following a
an
manualized protocol (Norton, 2012b). This protocol is designed to treat the common features of problematic anxiety rather than specific anxiety disorders by targeting broad vulnerabilities, such
M
as NA, rather than on specific features of a particular anxiety disorder. The facilitators in each
d
group were two doctoral student clinicians trained in CBT for anxiety that were supervised by
te
the second author. During the first component of treatment, the primary ingredients were psychoeducation, self-monitoring of anxiety, stress, and depression, cognitive restructuring of
Ac ce p
automatic thoughts specific to each individual’s referral concerns, and graduated exposure based on each individual’s fear hierarchy. The final component of treatment consists of cognitive restructuring focused more broadly on underlying schemas or core beliefs that underlie general manifestations of emotion symptoms. Every week, subjects completed the State version of the State-Trait Anxiety Inventory. After weeks 6 and 12, subjects completed the same questionnaire battery that they did at their initial assessment. Following the completion of the 12-week course of TGCBT, participants completed a second ADIS-IV. Results
Page 12 of 29
Mediators of Transdiagnostic CBT
12
The criterion variables, session-by-session STAI scores, were modeled over time using a mixed-effect linear regression using a Maximum Likelihood estimator. The data for the STAI, as well as the potential mediator variables, are shown in Figure 1. Results suggested good fit to the
ip t
data, RMSEA = 0.05, SRMR = 0.08, CFI = 0.95, with an average session 1 intercept of 47.76 (se = 0.59) and a significantly decreasing slope parameter (-1.01, se = 0.06, p < .001). Similarly, the
cr
potential mediator variables were modeled at pre-treatment, mid-treatment, and post-treatment
us
and showed adequate fit to the data, RMSEA = 0.09, SRMR = 0.08, CFI = 0.95. Each of the potential mediator variables showed expected estimated intercepts (PANAS-NA: 29.29, se =
an
0.49; ASI: 31.68, se = 0.76; IUS: 79.73, se = 1.56) and significantly decreasing slope parameters (PANAS-NA: -4.47, se = 0.32, p < .001; ASI: -6.56, se = 0.53, p < .001; IUS: -8.60, se = 1.07, p
M
< .001). Each of the slope parameters of the potential mediator variables showed positive simple
d
correlations with the STAI slope parameter, rs = 0.50 - 0.87, ps < .001, indicating that decreases
anxiety severity.
te
in each of the potential mediator variables were independently associated with decreases in
Ac ce p
Each of the slope parameters for the potential mediator variables were then regressed together onto the slope parameter of the STAI to identify the extent to which change on the each of the potential mediator variables uniquely predicts change in anxiety severity. Additionally, primary diagnosis was included in the model to identify possible differential change in anxiety severity by diagnosis, as were interaction terms of each potential mediator by diagnosis (i.e., PANAS-NA x Dx; ASI x Dx; IUS x Dx) to examine for differential relationships between change in potential mediators and change in anxiety severity by diagnosis. Results indicated a significant overall model, R2 = 0.84, F(11,256) = 123.67, p < .001. Inspection of the main effects of the potential mediator variables, revealed that change on the
Page 13 of 29
Mediators of Transdiagnostic CBT
13
PANAS-NA variable was positively, uniquely, and strongly related to change in STAI scores, r2 = 0. 74, B = 0.27 (Bootstrapped 95% CI: 0.22 – 0.35), p = .001, after accounting for variability associated with the other variables in the model. Neither change on ASI nor IUS showed
ip t
significant unique relationships with change in STAI, ASI: r2 = 0.05, B = -0.02 (Bootstrapped 95% CI: -0.05 – 0.03), p = .27; IUS: r2 = 0.05, B = -0.02 (Bootstrapped 95% CI: -0.05 – 0.003), p
cr
= .20, after controlling for the other variables in the model. Similarly, the main effect of primary
us
diagnosis, r2 < 0.001, p = .98, was not significant, indicating no differential change in anxiety severity among participants with primary diagnoses of panic disorder, social phobia, or GAD.
an
Finally, none of the potential mediator by diagnosis interaction terms was significant, r2s < 0.01, p > .36, suggesting no differential association between potential mediators and change in anxiety
M
severity among participants with primary diagnoses of panic disorder, social phobia, or GAD.
d
Discussion
te
The aim of this study was to examine mechanisms of change in transdiagnostic group CBT for anxiety. The results of the study indicated that change in anxiety severity throughout a
Ac ce p
12-week transdiagnostic CBGT program was best accounted for by change in NA during treatment. Although changes in AS and IU showed expected strong positive zero-order relationships with change in anxiety severity, after controlling for shared variability with change in NA, neither the AS and IU change scores were significantly associated with change in anxiety severity. Interestingly, the relationship of change in NA, AS, and IU showed no differential association with change in anxiety severity among participants with primary diagnoses of PD, SAD, and GAD, suggesting that the potential mediational role of NA is robust across diagnoses. Many transdiagnostic models of anxiety are predicated on the common underlying role of NA across diagnoses (Barlow, Allen, & Choate, 2004; Norton, 2006), and the current results
Page 14 of 29
Mediators of Transdiagnostic CBT
14
suggest that transdiagnostic CBT based on these models may achieve its effects via the modification of the underlying NA. Although some studies have suggested that the relationship of NA on anxiety may be differentially mediated by AS or IU for specific anxiety disorders (e.g.,
ip t
Sexton et al., 2003; Norton et al., 2005; Norton & Mehta, 2007), the current study found no
differential potential mediational role of change in AS or IU by diagnosis on change in anxiety
cr
severity during transdiagnostic CBT. This supports transdiagnostic theory, as therapeutic change
us
is more strongly associated with the higher-order risk factor than either of the second-order, specific risk factors. Despite the heterogeneous symptoms of anxiety disorders, this study
an
identified a homogeneous shared risk as a potential mechanism of change in anxiety treatment. Further, NA is implicated in a variety of psychological syndromes, such as personality disorders,
M
eating disorders, substance use disorders, and somatoform disorders (Mineka, Watson, & Clark,
te
warrants further research.
d
1998). Thus, transdiagnostic CBT has implications for a broad range of disorders and its utility
It is interesting that the majority of change in NA occurs during the first half of treatment
Ac ce p
when the focus is more on syndromal manifestations of NA, as opposed to the second half where the emphasis shifts to more global schema-level beliefs tied to NA. It may be that the core treatment elements of cognitive restructuring and exposure are sufficient to affect change in NA. Indeed, demonstrating that NA was the strongest predictor of changes in anxiety does not automatically indicate that treatment interventions directly targeting NA are responsible for those changes. As such, future dismantling studies should examine the relative short- and long-term impact of the later-session cognitive restructuring of underlying core beliefs on NA. This study tested for moderating effects due to primary diagnoses of PD, SAD, and GAD; as such, testing the mediational effects of these variables on other related diagnoses is warranted.
Page 15 of 29
Mediators of Transdiagnostic CBT
15
Further, including comorbidity in the analyses has implications for transdiagnostic anxiety treatment. It is possible that there may be different effects on NA, AS, or IU depending on the presence and type of comorbid diagnoses, and the precedence of each. However, including
ip t
comorbid cases in this sample may have attenuated in the ability to observe change in AS and IU. Single-diagnosis cases may display more specificity in their relationships to AS and IU.
cr
There were several limitations to this study. Although STAI data was collected weekly,
us
the PANAS, ASI, and IUS were only administered pre-treatment, mid-treatment (week 6 of 12), and post-treatment. This further reduces this study’s ability to establish temporal precedence. A
an
comparative lack of data points may lower the reliability of the observed effect. More data points would improve our ability to measure the trends with precision. Additionally, there are other
M
potential mediators, such as emotion dysregulation (e.g., Berking et al., 2008) or rumination
d
(McLaughlin & Nolen-Hoeksema, 2011), besides those used in this investigation. Future
te
research should use similar procedures to this study to determine if there are other potential mediators. This study also lacked a comparison group. It is possible that transdiagnostic CBT
Ac ce p
may operate via different mechanisms than diagnosis-specific CBT. Similarly, the use of comorbid cases may have reduced statistical power to detect specific effects of AS and IU. One could argue that the heterogeneity of this sample reduced the ability to observe evidence for second-order factor specificity to a particular diagnosis and that the tests of interactions conducted in this study provided only provided modest evidence to refute this. This evidence may also be attenuated by the large number of variables used in the regression analyses, which reduced power to detect these effects. Further, this study lacked the power to compare individuals with a single diagnosis against those with comorbid diagnoses. Based on the findings of McTeague and Lang (2012), there may be differences between these two groups in the role of
Page 16 of 29
Mediators of Transdiagnostic CBT
16
cognitive vulnerabilities in an individual presentation of anxiety, and therefore there may be value in comparing them. Future large scales studies should seek to replicate these findings with larger sample sizes of each diagnostic category.
ip t
While the current study is limited by the inability to establish temporal precedence of the potential mediators on change in anxiety, it does help establish that NA as the most promising
cr
overarching factor for future temporally-controlled mediational change-mechanism studies of
us
transdiagnostic CBT for anxiety disorder (see Kazdin, 2007, 2009). Given the prevalence of comorbidity in anxiety treatment, clinicians must be able to address multiple concerns with
an
clients. As NA is a core risk in both internalizing and externalizing disorders (Mineka, Watson, & Clark, 1998), transdiagnostic CBT appears to be a parsimonious method for treating processes
Ac ce p
te
d
M
inherent to all anxiety disorders and not simply specific diagnoses.
Page 17 of 29
Mediators of Transdiagnostic CBT
17 References
Arch, J.J., Wolitzky-Taylor, K.B., Eifert, G.H., & Craske, M.G. (2012). Longitudinal treatment mediation of traditional cognitive behavioral therapy and acceptance and commitment
ip t
therapy for anxiety disorders. Behaviour Research and Therapy, 50, 469-478. doi:10.1016/j.brat.2012.04.007
cr
Barlow, D. H. (2000). Unraveling the mysteries of anxiety and its disorders from the perspective
us
of emotion theory. American Psychologist, 55, 1247-1263. doi:10.1037/0003066X.55.11.1247
an
Barlow, D.H., Allen, L.B., & Choate, M.L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35. 205-230. doi:10.1016/S0005-7894(04)80036-4
M
Barlow, D. H., Gorman, J. M., Shear, M., & Woods, S. W. (2000). Cognitive-behavioral therapy,
d
imipramine, or their combination for panic disorder: A randomized controlled trial.
te
JAMA: Journal of the American Medical Association, 283, 2529-2536. doi:10.1001/jama.283.19.2529
Ac ce p
Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. doi:10.1037/0022-3514.51.6.1173 Belloch, A., Cabedo, E., Carrió, C., Fernández-Alvarez, H., García, F., Larsson, C. (2011). Group versus individual cognitive treatment for Obsessive-Complusive Disorder: Changes in non-OCD symptoms and cognitions at post-treatment and one-year followup. Psychiatry Research, 187, 174-179. doi:10.1016/j.psychres.2010.10.015
Page 18 of 29
Mediators of Transdiagnostic CBT
18
Berking, M., Wupperman, P., Reichardt, A., Pejic, T., Dippel, A., & Znoj, H. (2008). Emotionregulation skills as a treatment target in psychotherapy. Behaviour Research and Therapy, 46, 1230-1237. doi:10.1016/j.brat.2008.08.005
ip t
Boswell, J. F., Farchione, T. J., Sauer-Zavala, S., Murray, H. W., Fortune, M. R., & Barlow, D. H. (2013). Anxiety sensitivity and interoceptive exposure: A transdiagnostic construct
cr
and change strategy. Behavior Therapy, 44(3), 417-431. doi:10.1016/j.beth.2013.03.006
us
Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative
an
affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179-192. doi:10.1037/0021-843X.107.2.179
M
Brown, T. A., Di Nardo, P. A., & Barlow, D. H. (1994). Anxiety disorders interview schedule for
d
DSM-IV (Adult Version). Albany, NY: Graywind.
te
Buhr, K. K., & Dugas, M. J. (2002). The Intolerance of Uncertainty Scale: Psychometric properties of the English version. Behaviour Research and Therapy, 40, 931-946.
Ac ce p
doi:10.1016/S0005-7967(01)00092-4 Carleton, R., Mulvogue, M. K., Thibodeau, M. A., McCabe, R. E., Antony, M. M., & Asmundson, G. G. (2012). Increasingly certain about uncertainty: Intolerance of uncertainty across anxiety and depression. Journal of Anxiety Disorders, 26, 468-479. doi:10.1016/j.janxdis.2012.01.011 Chamberlain, L.D., & Norton, P.J. (2013). An evaluation of the effects of diagnostic composition on individual treatment outcome within transdiagnostic cognitive-behavioral group therapy for anxiety. Cognitive Behaviour Therapy, 42, 56-63. doi:10.1080/16506073.2012.748090
Page 19 of 29
Mediators of Transdiagnostic CBT
19
Clark, D.A. (2009). Cognitive behavioral therapy for anxiety and depression: Possibilities and limitations of a transdiagnostic perspective. Cognitive Behaviour Therapy, 38(S1), 29-34. doi:10.1080/16506070992980745
ip t
Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal Of Abnormal Psychology, 100, 316-336.
cr
doi:10.1037/0021-843X.100.3.316
us
Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology, 103, 103-116. doi:10.1037/0021-
an
843X.103.1.103
Covin, R., Ouimet, A. J., Seeds, P. M., & Dozois, D. A. (2008). A meta-analysis of CBT for
d
doi:10.1016/j.janxdis.2007.01.002
M
pathological worry among clients with GAD. Journal of Anxiety Disorders, 22, 108-116.
te
Dugas, M. J., Gagnon, F., Ladouceur, R., & Freeston, M. H. (1998). Generalized anxiety disorder: A preliminary test of a conceptual model. Behaviour Research and Therapy, 36,
Ac ce p
215–226. doi:10.1016/S0005-7967(97)00070-3 Dugas, M. J., Ladouceur, R., Léger, E., Freeston, M. H., Langolis, F., Provencher, M. D., & Boisvert, J. (2003). Group cognitive-behavioral therapy for generalized anxiety disorder: Treatment outcome and long-term follow-up. Journal of Consulting and Clinical Psychology, 71, 821-825. doi:10.1037/0022-006X.71.4.821 Fedoroff, I.C., Taylor, S., Asmundson, G.J., & Koch, W.J. (2000). Cognitive factors in traumatic stress reactions: Predicting PTSD symptoms from anxiety sensitivity and beliefs about harmful events. Behavioural and Cognitive Psychotherapy, 28, 5-15.
Page 20 of 29
Mediators of Transdiagnostic CBT
20
Fisher, P.L, & Durham, R.C. (1999). Recovery rates in generalized anxiety disorder following psychological therapy: An analysis of clinically significant change in the SATIT across outcome studies since 1990. Psychological Medicine, 29, 1425-1434.
ip t
doi:10.1017/S0033291799001336
Freeston, M.H., Rhéaume, J., Letarte, H., Dugas, M.J., & Ladouceur, R. (1994). Why do people
cr
worry? Personality and Individual Differences, 17, 791-802. doi:10.1037/0022-
us
006X.65.3.405
Hayes, A.F. (2009). Beyond Baron and Kenney: Statistical mediation analysis in the new
an
millennium. Communication Monographs, 76(4), 408-420. Doi:10.1080/03637750903310360
M
Hazen, A. L., Walker, J. R., & Eldridge, G. D. (1996). Anxiety sensitivity and treatment outcome
te
ANXI5>3.0.CO;2-D
d
in panic disorder. Anxiety, 2, 34-39. doi:10.1002/(SICI)1522-7154(1996)2:1
S0887-6185(14)00141-8 http://dx.doi.org/doi:10.1016/j.janxdis.2014.09.017 ANXDIS 1652
To appear in:
Journal of Anxiety Disorders
Received date: Revised date: Accepted date:
30-4-2014 15-9-2014 26-9-2014
Please cite this article as: Talkovsky, A. M., and Norton, P. J.,Mediators of Transdiagnostic Group Cognitive Behavior Therapy, Journal of Anxiety Disorders (2014), http://dx.doi.org/10.1016/j.janxdis.2014.09.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Peter J. Norton, Ph.D. 1 University of Houston
Ac ce p
te
d
M
an
us
1
cr
Alexander M. Talkovsky, M.A. 1
ip t
Mediators of Transdiagnostic Group Cognitive Behavior Therapy
Page 1 of 29
RUNNING HEAD: Mediators of Transdiagnostic CBT
1
Journal of Anxiety Disorders
an
us
September 15, 2014
cr
Submitted to:
ip t
Mediators of Transdiagnostic Group Cognitive Behavior Therapy
Word Count: 5804
Ac ce p
te
d
M
Tables and Figures: 1
Page 2 of 29
Mediators of Transdiagnostic CBT
2 Abstract
The efficacy of cognitive-behavioral therapy (CBT) for anxiety is well established. Investigations into the mechanisms of change in CBT report changes in cognitive vulnerabilities
ip t
mediating improvements over the course of treatment. As anxiety disorders share certain risk factors, there is a trend towards CBT emphasizing these vulnerabilities, including negative
cr
affectivity (NA) and also more specific constructs such as anxiety sensitivity (AS) and
us
intolerance of uncertainty (IU). The purpose of this investigation was to analyze potential mediators of anxiety reduction over the course of transdiagnostic group CBT. NA, AS, and IU all
an
decreased over the course of treatment. Among the potential mediators, change in NA had a significant relationship with change in anxiety but change in AS and change in IU did not.
M
Neither the main effect of primary diagnosis nor the interactions between potential mediators and
d
primary diagnoses were significant, indicating that there were no differential changes in anxiety
te
or the potential mediators across primary diagnoses. Results strongly point towards NA as an
Ac ce p
overarching mediator of anxiety reduction during transdiagnostic group CBT.
Keywords: anxiety, transdiagnostic, cognitive behavior therapy, mediation
Page 3 of 29
Mediators of Transdiagnostic CBT
3
Mediators of Transdiagnostic Group Cognitive Behavior Therapy The most common class of psychological disorders is anxiety, as roughly a quarter to a third of all people will meet criteria for an anxiety disorder in their lifetimes (Kessler et al., 2005;
ip t
Kessler et al., 2012). A large body of research supports the use of cognitive behavioral therapy (CBT) for the treatment of anxiety (Norton & Price, 2007). CBT-based interventions have robust
cr
effects on panic (Barlow, Gorman, Shear, & Woods, 2000), social phobia (Heimberg, 2002)
us
generalized anxiety disorder (GAD; Covin, Ouimet, Seeds, & Dozois, 2008), posttraumatic stress disorder (PTSD; Powers et al., 2010), and obsessive compulsive disorder (OCD; Whittal,
an
Thordarson, & McLean, 2005), and clinically significant gains are maintained long-term. Despite the large body of empirical support for CBT, practitioners often struggle to
M
deliver care to all those that would benefit from it (Shafran et al., 2009), partially because it is
d
time-consuming, costly, and inefficient to disseminate multiple disorder-specific manualized
te
CBT protocols (Clark, 2009). There is a trend towards anxiety treatment that emphasizes vulnerability shared across the anxiety disorders, as there is a large amount of overlap (Barlow,
Ac ce p
2000; Norton, 2006). In Clark and Watson’s (1991) tripartite model of anxiety and depression, depression and anxiety share the underlying vulnerability of negative affectivity, or a dispositional penchant for responding to negative stimuli with feelings of fear, anxiety, depression, guilt and self-dissatisfaction (Clark, Watson, & Mineka, 2004). Negative affectivity serves as a higher-order factor subsuming lower-order risk factors that confer more specific risk for certain disorders. Brown, Chorpita, and Barlow (1998) observed a model in which negative affectivity had a direct causal influence on GAD, panic disorder, OCD, and social phobia and this model had an excellent fit to their data. Watson (2005) suggested that a model of anxiety disorders should not only include factors that cut across diagnoses (e.g., negative affectivity), but
Page 4 of 29
Mediators of Transdiagnostic CBT
4
also those that are more specific to particular conditions. The most heavily studied constructs proposed to be specific factors may be anxiety sensitivity and intolerance of uncertainty. Anxiety sensitivity, or the fear of fearful and/or anxious symptoms due to the belief that
ip t
they may have adverse consequences (Reiss & McNally, 1985) is heavily implicated in panic disorder (McNally, 2002) and PTSD (Taylor, 1993), but scores are also elevated among
cr
individuals with GAD, social phobia, and OCD compared to non-anxious controls (Taylor,
us
Koch, & McNally, 1992). Furthermore, Hazen, Walker, and Eldridge (1996) found that anxiety sensitivity (AS) is responsive to anxiety treatment, and this improvement is related to the
an
reduction of anxiety symptoms. In motor vehicle accident victims, Fedoroff, Taylor, Asmundson, and Koch (2000) found, using regression analyses, that AS and pain severity were significant
M
predictors of both trauma symptoms prior to treatment and symptom reduction following
d
treatment. Smits, Powers, Cho, and Telch (2004) used the Baron and Kenny (1986) method to
te
test whether a reduction in fear of fear underlies improvement resulting from CBT for panic disorder. They observed full mediation of the relationship between the effects of CBT and
Ac ce p
changes in global impairment, and partial mediation on agoraphobia, anxiety, and panic frequency. Hofmann and colleagues (2007) substantiated these results, finding that panic-related catastrophic cognitions mediated treatment change in CBT. Few studies analyze the role of anxiety sensitivity in treatment of other anxiety disorders, but rather they focus on panic disorder (Smits, Powers, Cho, & Telch, 2004) or PTSD (Fedoroff, Taylor, Asmundson, & Koch, 2000), leaving a need to analyze the rest of the anxiety disorders. Boswell and colleagues (2013) observed elevated AS in patients with primary diagnoses of panic disorder, social phobia, GAD, and OCD, and all groups demonstrated AS improvements following interoceptive exposures, and lower levels of AS were associated with lower levels of
Page 5 of 29
Mediators of Transdiagnostic CBT
5
clinical severity in analyses collapsed across principal diagnoses. Thus, AS may represent a specific vulnerability to panic, but it may confer risk to the other anxiety disorders as well. Similarly, the construct of intolerance of uncertainty (IU) has been operationalized in
ip t
many different ways (Starcevic & Berle, 2006). For the purposes of this investigation, it will be defined as a characteristic tendency to find uncertainty regarding outcomes of an event aversive
cr
(Freeston, Rhéaume, Letarte, Dugas, & Ladouceur, 1994). Treatment specifically targeting IU in
us
individuals with GAD leads to clinically significant change immediately following treatment and gains are maintained at twelve-month follow up whether treating individuals (Ladouceur et al.,
an
2000) or groups (Dugas et al., 2003). Although it was originally hypothesized that IU is specific to GAD (Dugas, Gagnon, Ladouceur, & Freeston, 1998), Carleton and colleagues (2012) found
M
elevated scores in social anxiety disorder (SAD), panic disorder (PD), GAD, OCD, and
d
depression patients, and concluded that IU may represent a transdiagnostic vulnerability.
te
McEvoy and Mahoney (2012) found that IU mediates the relationship between neuroticism, a higher order vulnerability, and social anxiety, panic, and depression; also, they found that the
Ac ce p
prospective anxiety factor of IU mediated the relationship between neuroticism and worry, and obsessive-compulsive symptoms. Belloch and colleagues (2011) found that IU decreased over the course of cognitive therapy for OCD. Mahoney and McEvoy (2012) observed reductions in IU specific to social situations following group CBT for SAD, but it is possible that this situation-specific assessment captures variance overlapping with diagnostic criteria for SAD and it would be worth replicating results with a more general assessment tool. IU is clearly relevant to the treatment process, but presently lacks empirical support as a mediator of treatment outcomes for anxiety. Despite its associations with GAD, IU is a promising transdiagnostic marker for anxiety in light of its associations with a number of other disorders.
Page 6 of 29
Mediators of Transdiagnostic CBT
6
Sexton, Norton, Walker, and Norton (2003) tested a hierarchical model in a sample of undergraduates with neuroticism serving as the first-order factor subsuming AS and IU. They found that AS acted as a risk for panic, obsessive-compulsive symptoms, and health anxiety, IU
ip t
acted as a risk for worry, and neuroticism was an overarching risk for all symptom types. Norton, Sexton, Walker, and Norton (2005) partially replicated these results in a sample of treatment-
cr
seekers. They observed that negative affectivity served as a global risk factor and influenced AS
us
and IU, AS affected panic, and health anxiety, and IU influenced worry and depression; however, the expected relationship between negative affectivity and obsessive-compulsive
an
symptoms was not observed. Norton and Mehta (2007) used structural equation modeling to compare potential hierarchical models. They improved the fit of previous models by including a
M
path from IU to obsessive-compulsive and social anxiety symptoms. These studies combine to
d
provide support for hierarchical models of anxiety with AS and IU acting as second-order risk
te
factors that show a degree of specificity, but not as much as some originally thought. As IU and AS are oft-acknowledged risk factors, and they decrease across treatment, it
Ac ce p
would be enlightening to determine their relationships to symptom reduction and to do so across diagnoses. Because of their demonstrated relationships with anxiety disorders and their response to treatment, this study posits AS and IU as potential mediators of anxiety reduction. Mediator variables explain the way an antecedent exerts its effects on a dependent variable, thus allowing a more complete understanding of how the criterion and outcome variables relate to one another (MacKinnon & Fairchild, 2009). Mediational analyses are important for treatment and prevention, as interventions are often designed to change an outcome by targeting mediators hypothesized to have a causal relationship with the outcome (MacKinnon, Fairchild, & Fritz, 2007). It is important to understand what makes therapy work because understanding
Page 7 of 29
Mediators of Transdiagnostic CBT
7
mechanisms of change can improve the identification of appropriate, better, different, and new treatments (Kazdin, 2007, 2009). Much of the extant literature suggests that cognitive variables mediate the outcomes of diagnosis-specific anxiety treatment (Arch, Wolitzky-Taylor, Eifert, &
ip t
Craske, 2012; Hofmann, 2004; Hofmann et al., 2007; Mahoney & McEvoy, 2012; Smits, Powers, Cho, & Telch, 2004).
cr
As interest in transdiagnostic psychotherapy grows, it is important to establish the role of
us
shared vulnerabilities, such as AS and IU, in the course of psychotherapy. In a mixed anxiety disorder sample of 35 individuals randomized to receive CBT that included components shared
an
across the anxiety disorders, Arch, Wolitzky-Taylor, Eifert, and Craske (2012) found that AS significantly mediated post-treatment worry and they observed no differential effects by
M
diagnosis. However, they failed to observe significant mediation for four of their five outcome
d
measures. Updating these results with a larger sample would contribute to the growing body of
te
transdiagnostic literature. We attempt to improve upon cross-sectional analysis by investigating simultaneous change over the entire course of a manualized treatment protocol. This study aims
Ac ce p
to identify potential mediators by observing the change in potential criterion, mediator, and outcome variables.
The purpose of this study is to analyze potential mediators of anxious symptom reduction in transdiagnostic cognitive behavioral group therapy. Norton (2008) established the effectiveness of transdiagnostic CBGT, but the mechanisms of change have not yet been established. This study seeks to identify potential mediators of overall anxiety reduction across anxiety diagnoses. The first hypothesis is that reductions in negative affectivity, a higher-order risk for anxiety, will be a significant mediator of symptom improvement across treatment in all diagnostic categories. Second, it is hypothesized that AS and IU will each act as specific
Page 8 of 29
Mediators of Transdiagnostic CBT
8
mediators, with AS significantly mediating improvement in panic symptoms and social anxiety, and IU mediating the improvement in worry. Methods
ip t
The data used in this study was collected during three previous clinical trials (Norton, 2008, 2012a; Norton & Barrera, 2012) that share methodologies and treatment protocols
cr
(Norton, 2012b). These results have not been previously published. All procedures were
us
approved for human subjects research by the Institutional Review Board at the University of Houston.
an
Participants
Participants included 256 individuals with principal diagnoses of panic disorder (PD),
M
social phobia (SP) or generalized anxiety disorder (GAD) seeking treatment at the Anxiety
d
Disorder Clinic (ADC) at the University of Houston. All of the individuals included in these
te
analyses were randomized into transdiagnostic cognitive behavioral group therapy. Over half (52.2%) of the sample identified as female. The mean age was 32.89 years (S.D. = 10.72). The
Ac ce p
sample was racially diverse; 56.1% self-identified as Caucasian, 9.6% as African-American, 7.4% as Asian or Asian-American, 20.7% as Hispanic or Latino(a), and 6.3% self-identified as multiracial or as a member as “Other.” The majority of the sample (57.7%) reported having completed at least a bachelor’s degree, 36.5% reported that they are married or cohabitating, and 76.2% reported that they are employed full time or are enrolled as full-time students. Income data were not collected. In this sample, 131 individuals (51.1%) had a primary diagnosis of SP. Among them, 31 (23.7%) had comorbid GAD, and 8 had comorbid PD (6.1%). Seventy eight individuals (30.5%) had a primary diagnosis of PD, and among them, 13 (16.7%) had comorbid SP and 24 (30.8%)
Page 9 of 29
Mediators of Transdiagnostic CBT
9
had comorbid GAD. Forty seven (18.4%) individuals had a primary diagnosis of GAD; 12 (25.5%) of them also had a diagnosis of PD, and 21 (44.7%) had a comorbid SP diagnosis. Comorbid depressive diagnoses were relatively common, but did not differ in frequency across
ip t
individuals with principal diagnoses of PD (32.1%), SP (27.5%), or GAD (44.7%), χ2 (2, n = 256) = 4.70, p = .10.
cr
Measures
us
Anxiety Disorders Interview Schedule for DSM-IV (ADIS-IV). The ADIS-IV (Brown, Di Nardo, & Barlow, 1994) is a semi-structured diagnostic interview that assess for the presence
an
and severity of anxiety, mood, somatoform; it also provides a brief screen for psychotic symptoms and substance abuse and dependence. Previous research on this sample reported high
M
interrater reliability and diagnostic agreement (86% agreement, κ = 0.759, p < .001; Chamberlain
d
& Norton, 2013).
te
State-Trait Anxiety Inventory – State Version (STAI-S). The STAI (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1993) is a 20-item self-report measure designed to assess
Ac ce p
state and trait levels of anxiety. Only the state version was used for the purposes of this investigation. Items are scored from 1 (“not at all”) to 4 (“very much so”) to indicate how much each statement reflects the subject’s present feelings; total scores range from 20 to 80. The original manual reported strong psychometric properties in several populations. Fisher and Durham (1999) reported that it is sensitive to treatment effects. In this sample, its internal consistency was excellent at session 1 (Cronbach α = .929). Positive and Negative Affect Schedule (PANAS). The PANAS (Watson, Clark, & Tellegen, 1988) is a 20-item self-report measure designed to assess positive affect (PA) and negative affect (NA). PA and NA are considered orthogonal measures of independent affective
Page 10 of 29
Mediators of Transdiagnostic CBT
10
dimensions (Watson, Clark, & Carey, 1988), but weak correlations are sometimes observed (Vittengl & Holt, 1998). This measure can assess either state or trait affect depending on the specific phrasing of the instructions (Watson, Clark, & Tellegen, 1988). For this study, the
ip t
instructions were phrased to measure trait affectivity. The PANAS has demonstrated acceptable psychometric properties in clinically anxious and depressed samples (Brown et al., 1998;
cr
Watson, Clark, & Carey, 1988), and good reliability in the current sample at pre-treatment
us
(Cronbach α = .821).
Anxiety Sensitivity Index (ASI). The ASI (Reiss, Peterson, Gursky, & McNally, 1986;
an
Peterson & Reiss, 1993) is a 16-item self-report measure designed to assess the fear of anxietyrelated symptoms. Items are scored from 0 (very little) to 4 (very much). Individuals with
M
anxiety disorder diagnoses score significantly higher than non-clinical controls, although the ASI
d
shows greater specificity for panic (Taylor, Koch, & McNally, 1992). It demonstrates adequate
te
criterion and construct validities (Reiss et al., 1986; Peterson & Reiss, 1993). The ASI demonstrated good internal consistency (Cronbach α = .865) at pre-treatment in this sample.
Ac ce p
Intolerance of Uncertainty Scale (IUS). The IUS (Freeston, Rhéaume, Letarte, Dugas, & Ladouceur, 1994) is a 27-item self-report measure designed to assess the extent to which an individual finds uncertainty to be aversive. It demonstrates adequate convergent and discriminant validity (Buhr & Dugas, 2002). The IUS demonstrated excellent internal consistency at pretreatment (Cronbach α = .957) in this sample. Procedures Prior to assessment, subjects provided their informed consent to participate in research and to receive treatment through the ADC. Then, they completed a large battery of self-report assessments, including the STAI, PANAS, ASI, and IUS as well as other measures not reltead to
Page 11 of 29
Mediators of Transdiagnostic CBT
11
this study. Next, subjects completed the ADIS-IV as administered by a trained graduate student and were subsequently provided a copy of their assessment report. Individuals with anxiety disorders that did not demonstrate dementia or other serious cognitive decline, adequately spoke
ip t
and read English, and did not display serious suicidality, substance use, or other conditions
requiring immediate intervention were invited to participate in a transdiagnostic group CBT
cr
program. The transdiagnostic model focuses on a global formulation of dysfunctional anxiety
us
rather than diagnosis-specific conceptualization.
The treatment program consisted of 12 weekly 2-hour long sessions following a
an
manualized protocol (Norton, 2012b). This protocol is designed to treat the common features of problematic anxiety rather than specific anxiety disorders by targeting broad vulnerabilities, such
M
as NA, rather than on specific features of a particular anxiety disorder. The facilitators in each
d
group were two doctoral student clinicians trained in CBT for anxiety that were supervised by
te
the second author. During the first component of treatment, the primary ingredients were psychoeducation, self-monitoring of anxiety, stress, and depression, cognitive restructuring of
Ac ce p
automatic thoughts specific to each individual’s referral concerns, and graduated exposure based on each individual’s fear hierarchy. The final component of treatment consists of cognitive restructuring focused more broadly on underlying schemas or core beliefs that underlie general manifestations of emotion symptoms. Every week, subjects completed the State version of the State-Trait Anxiety Inventory. After weeks 6 and 12, subjects completed the same questionnaire battery that they did at their initial assessment. Following the completion of the 12-week course of TGCBT, participants completed a second ADIS-IV. Results
Page 12 of 29
Mediators of Transdiagnostic CBT
12
The criterion variables, session-by-session STAI scores, were modeled over time using a mixed-effect linear regression using a Maximum Likelihood estimator. The data for the STAI, as well as the potential mediator variables, are shown in Figure 1. Results suggested good fit to the
ip t
data, RMSEA = 0.05, SRMR = 0.08, CFI = 0.95, with an average session 1 intercept of 47.76 (se = 0.59) and a significantly decreasing slope parameter (-1.01, se = 0.06, p < .001). Similarly, the
cr
potential mediator variables were modeled at pre-treatment, mid-treatment, and post-treatment
us
and showed adequate fit to the data, RMSEA = 0.09, SRMR = 0.08, CFI = 0.95. Each of the potential mediator variables showed expected estimated intercepts (PANAS-NA: 29.29, se =
an
0.49; ASI: 31.68, se = 0.76; IUS: 79.73, se = 1.56) and significantly decreasing slope parameters (PANAS-NA: -4.47, se = 0.32, p < .001; ASI: -6.56, se = 0.53, p < .001; IUS: -8.60, se = 1.07, p
M
< .001). Each of the slope parameters of the potential mediator variables showed positive simple
d
correlations with the STAI slope parameter, rs = 0.50 - 0.87, ps < .001, indicating that decreases
anxiety severity.
te
in each of the potential mediator variables were independently associated with decreases in
Ac ce p
Each of the slope parameters for the potential mediator variables were then regressed together onto the slope parameter of the STAI to identify the extent to which change on the each of the potential mediator variables uniquely predicts change in anxiety severity. Additionally, primary diagnosis was included in the model to identify possible differential change in anxiety severity by diagnosis, as were interaction terms of each potential mediator by diagnosis (i.e., PANAS-NA x Dx; ASI x Dx; IUS x Dx) to examine for differential relationships between change in potential mediators and change in anxiety severity by diagnosis. Results indicated a significant overall model, R2 = 0.84, F(11,256) = 123.67, p < .001. Inspection of the main effects of the potential mediator variables, revealed that change on the
Page 13 of 29
Mediators of Transdiagnostic CBT
13
PANAS-NA variable was positively, uniquely, and strongly related to change in STAI scores, r2 = 0. 74, B = 0.27 (Bootstrapped 95% CI: 0.22 – 0.35), p = .001, after accounting for variability associated with the other variables in the model. Neither change on ASI nor IUS showed
ip t
significant unique relationships with change in STAI, ASI: r2 = 0.05, B = -0.02 (Bootstrapped 95% CI: -0.05 – 0.03), p = .27; IUS: r2 = 0.05, B = -0.02 (Bootstrapped 95% CI: -0.05 – 0.003), p
cr
= .20, after controlling for the other variables in the model. Similarly, the main effect of primary
us
diagnosis, r2 < 0.001, p = .98, was not significant, indicating no differential change in anxiety severity among participants with primary diagnoses of panic disorder, social phobia, or GAD.
an
Finally, none of the potential mediator by diagnosis interaction terms was significant, r2s < 0.01, p > .36, suggesting no differential association between potential mediators and change in anxiety
M
severity among participants with primary diagnoses of panic disorder, social phobia, or GAD.
d
Discussion
te
The aim of this study was to examine mechanisms of change in transdiagnostic group CBT for anxiety. The results of the study indicated that change in anxiety severity throughout a
Ac ce p
12-week transdiagnostic CBGT program was best accounted for by change in NA during treatment. Although changes in AS and IU showed expected strong positive zero-order relationships with change in anxiety severity, after controlling for shared variability with change in NA, neither the AS and IU change scores were significantly associated with change in anxiety severity. Interestingly, the relationship of change in NA, AS, and IU showed no differential association with change in anxiety severity among participants with primary diagnoses of PD, SAD, and GAD, suggesting that the potential mediational role of NA is robust across diagnoses. Many transdiagnostic models of anxiety are predicated on the common underlying role of NA across diagnoses (Barlow, Allen, & Choate, 2004; Norton, 2006), and the current results
Page 14 of 29
Mediators of Transdiagnostic CBT
14
suggest that transdiagnostic CBT based on these models may achieve its effects via the modification of the underlying NA. Although some studies have suggested that the relationship of NA on anxiety may be differentially mediated by AS or IU for specific anxiety disorders (e.g.,
ip t
Sexton et al., 2003; Norton et al., 2005; Norton & Mehta, 2007), the current study found no
differential potential mediational role of change in AS or IU by diagnosis on change in anxiety
cr
severity during transdiagnostic CBT. This supports transdiagnostic theory, as therapeutic change
us
is more strongly associated with the higher-order risk factor than either of the second-order, specific risk factors. Despite the heterogeneous symptoms of anxiety disorders, this study
an
identified a homogeneous shared risk as a potential mechanism of change in anxiety treatment. Further, NA is implicated in a variety of psychological syndromes, such as personality disorders,
M
eating disorders, substance use disorders, and somatoform disorders (Mineka, Watson, & Clark,
te
warrants further research.
d
1998). Thus, transdiagnostic CBT has implications for a broad range of disorders and its utility
It is interesting that the majority of change in NA occurs during the first half of treatment
Ac ce p
when the focus is more on syndromal manifestations of NA, as opposed to the second half where the emphasis shifts to more global schema-level beliefs tied to NA. It may be that the core treatment elements of cognitive restructuring and exposure are sufficient to affect change in NA. Indeed, demonstrating that NA was the strongest predictor of changes in anxiety does not automatically indicate that treatment interventions directly targeting NA are responsible for those changes. As such, future dismantling studies should examine the relative short- and long-term impact of the later-session cognitive restructuring of underlying core beliefs on NA. This study tested for moderating effects due to primary diagnoses of PD, SAD, and GAD; as such, testing the mediational effects of these variables on other related diagnoses is warranted.
Page 15 of 29
Mediators of Transdiagnostic CBT
15
Further, including comorbidity in the analyses has implications for transdiagnostic anxiety treatment. It is possible that there may be different effects on NA, AS, or IU depending on the presence and type of comorbid diagnoses, and the precedence of each. However, including
ip t
comorbid cases in this sample may have attenuated in the ability to observe change in AS and IU. Single-diagnosis cases may display more specificity in their relationships to AS and IU.
cr
There were several limitations to this study. Although STAI data was collected weekly,
us
the PANAS, ASI, and IUS were only administered pre-treatment, mid-treatment (week 6 of 12), and post-treatment. This further reduces this study’s ability to establish temporal precedence. A
an
comparative lack of data points may lower the reliability of the observed effect. More data points would improve our ability to measure the trends with precision. Additionally, there are other
M
potential mediators, such as emotion dysregulation (e.g., Berking et al., 2008) or rumination
d
(McLaughlin & Nolen-Hoeksema, 2011), besides those used in this investigation. Future
te
research should use similar procedures to this study to determine if there are other potential mediators. This study also lacked a comparison group. It is possible that transdiagnostic CBT
Ac ce p
may operate via different mechanisms than diagnosis-specific CBT. Similarly, the use of comorbid cases may have reduced statistical power to detect specific effects of AS and IU. One could argue that the heterogeneity of this sample reduced the ability to observe evidence for second-order factor specificity to a particular diagnosis and that the tests of interactions conducted in this study provided only provided modest evidence to refute this. This evidence may also be attenuated by the large number of variables used in the regression analyses, which reduced power to detect these effects. Further, this study lacked the power to compare individuals with a single diagnosis against those with comorbid diagnoses. Based on the findings of McTeague and Lang (2012), there may be differences between these two groups in the role of
Page 16 of 29
Mediators of Transdiagnostic CBT
16
cognitive vulnerabilities in an individual presentation of anxiety, and therefore there may be value in comparing them. Future large scales studies should seek to replicate these findings with larger sample sizes of each diagnostic category.
ip t
While the current study is limited by the inability to establish temporal precedence of the potential mediators on change in anxiety, it does help establish that NA as the most promising
cr
overarching factor for future temporally-controlled mediational change-mechanism studies of
us
transdiagnostic CBT for anxiety disorder (see Kazdin, 2007, 2009). Given the prevalence of comorbidity in anxiety treatment, clinicians must be able to address multiple concerns with
an
clients. As NA is a core risk in both internalizing and externalizing disorders (Mineka, Watson, & Clark, 1998), transdiagnostic CBT appears to be a parsimonious method for treating processes
Ac ce p
te
d
M
inherent to all anxiety disorders and not simply specific diagnoses.
Page 17 of 29
Mediators of Transdiagnostic CBT
17 References
Arch, J.J., Wolitzky-Taylor, K.B., Eifert, G.H., & Craske, M.G. (2012). Longitudinal treatment mediation of traditional cognitive behavioral therapy and acceptance and commitment
ip t
therapy for anxiety disorders. Behaviour Research and Therapy, 50, 469-478. doi:10.1016/j.brat.2012.04.007
cr
Barlow, D. H. (2000). Unraveling the mysteries of anxiety and its disorders from the perspective
us
of emotion theory. American Psychologist, 55, 1247-1263. doi:10.1037/0003066X.55.11.1247
an
Barlow, D.H., Allen, L.B., & Choate, M.L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35. 205-230. doi:10.1016/S0005-7894(04)80036-4
M
Barlow, D. H., Gorman, J. M., Shear, M., & Woods, S. W. (2000). Cognitive-behavioral therapy,
d
imipramine, or their combination for panic disorder: A randomized controlled trial.
te
JAMA: Journal of the American Medical Association, 283, 2529-2536. doi:10.1001/jama.283.19.2529
Ac ce p
Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. doi:10.1037/0022-3514.51.6.1173 Belloch, A., Cabedo, E., Carrió, C., Fernández-Alvarez, H., García, F., Larsson, C. (2011). Group versus individual cognitive treatment for Obsessive-Complusive Disorder: Changes in non-OCD symptoms and cognitions at post-treatment and one-year followup. Psychiatry Research, 187, 174-179. doi:10.1016/j.psychres.2010.10.015
Page 18 of 29
Mediators of Transdiagnostic CBT
18
Berking, M., Wupperman, P., Reichardt, A., Pejic, T., Dippel, A., & Znoj, H. (2008). Emotionregulation skills as a treatment target in psychotherapy. Behaviour Research and Therapy, 46, 1230-1237. doi:10.1016/j.brat.2008.08.005
ip t
Boswell, J. F., Farchione, T. J., Sauer-Zavala, S., Murray, H. W., Fortune, M. R., & Barlow, D. H. (2013). Anxiety sensitivity and interoceptive exposure: A transdiagnostic construct
cr
and change strategy. Behavior Therapy, 44(3), 417-431. doi:10.1016/j.beth.2013.03.006
us
Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative
an
affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179-192. doi:10.1037/0021-843X.107.2.179
M
Brown, T. A., Di Nardo, P. A., & Barlow, D. H. (1994). Anxiety disorders interview schedule for
d
DSM-IV (Adult Version). Albany, NY: Graywind.
te
Buhr, K. K., & Dugas, M. J. (2002). The Intolerance of Uncertainty Scale: Psychometric properties of the English version. Behaviour Research and Therapy, 40, 931-946.
Ac ce p
doi:10.1016/S0005-7967(01)00092-4 Carleton, R., Mulvogue, M. K., Thibodeau, M. A., McCabe, R. E., Antony, M. M., & Asmundson, G. G. (2012). Increasingly certain about uncertainty: Intolerance of uncertainty across anxiety and depression. Journal of Anxiety Disorders, 26, 468-479. doi:10.1016/j.janxdis.2012.01.011 Chamberlain, L.D., & Norton, P.J. (2013). An evaluation of the effects of diagnostic composition on individual treatment outcome within transdiagnostic cognitive-behavioral group therapy for anxiety. Cognitive Behaviour Therapy, 42, 56-63. doi:10.1080/16506073.2012.748090
Page 19 of 29
Mediators of Transdiagnostic CBT
19
Clark, D.A. (2009). Cognitive behavioral therapy for anxiety and depression: Possibilities and limitations of a transdiagnostic perspective. Cognitive Behaviour Therapy, 38(S1), 29-34. doi:10.1080/16506070992980745
ip t
Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal Of Abnormal Psychology, 100, 316-336.
cr
doi:10.1037/0021-843X.100.3.316
us
Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology, 103, 103-116. doi:10.1037/0021-
an
843X.103.1.103
Covin, R., Ouimet, A. J., Seeds, P. M., & Dozois, D. A. (2008). A meta-analysis of CBT for
d
doi:10.1016/j.janxdis.2007.01.002
M
pathological worry among clients with GAD. Journal of Anxiety Disorders, 22, 108-116.
te
Dugas, M. J., Gagnon, F., Ladouceur, R., & Freeston, M. H. (1998). Generalized anxiety disorder: A preliminary test of a conceptual model. Behaviour Research and Therapy, 36,
Ac ce p
215–226. doi:10.1016/S0005-7967(97)00070-3 Dugas, M. J., Ladouceur, R., Léger, E., Freeston, M. H., Langolis, F., Provencher, M. D., & Boisvert, J. (2003). Group cognitive-behavioral therapy for generalized anxiety disorder: Treatment outcome and long-term follow-up. Journal of Consulting and Clinical Psychology, 71, 821-825. doi:10.1037/0022-006X.71.4.821 Fedoroff, I.C., Taylor, S., Asmundson, G.J., & Koch, W.J. (2000). Cognitive factors in traumatic stress reactions: Predicting PTSD symptoms from anxiety sensitivity and beliefs about harmful events. Behavioural and Cognitive Psychotherapy, 28, 5-15.
Page 20 of 29
Mediators of Transdiagnostic CBT
20
Fisher, P.L, & Durham, R.C. (1999). Recovery rates in generalized anxiety disorder following psychological therapy: An analysis of clinically significant change in the SATIT across outcome studies since 1990. Psychological Medicine, 29, 1425-1434.
ip t
doi:10.1017/S0033291799001336
Freeston, M.H., Rhéaume, J., Letarte, H., Dugas, M.J., & Ladouceur, R. (1994). Why do people
cr
worry? Personality and Individual Differences, 17, 791-802. doi:10.1037/0022-
us
006X.65.3.405
Hayes, A.F. (2009). Beyond Baron and Kenney: Statistical mediation analysis in the new
an
millennium. Communication Monographs, 76(4), 408-420. Doi:10.1080/03637750903310360
M
Hazen, A. L., Walker, J. R., & Eldridge, G. D. (1996). Anxiety sensitivity and treatment outcome
te
ANXI5>3.0.CO;2-D
d
in panic disorder. Anxiety, 2, 34-39. doi:10.1002/(SICI)1522-7154(1996)2:1
