Reductions in negative repetitive thinking and metacognitive beliefs

Behaviour Research and Therapy 59 (2014) 52e60
Contents lists available at ScienceDirect
Behaviour Research and Therapy
journal homepage: www.elsevier.com/locate/brat
Reductions in negative repetitive thinking and metacognitive beliefs
during transdiagnostic internet cognitive behavioural therapy (iCBT)
for mixed anxiety and depression
Jill M. Newby*, Alishia D. Williams, Gavin Andrews
Clinical Research Unit for Anxiety and Depression (CRUfAD), School of Psychiatry, The University of New South Wales (UNSW) at St. Vincent's Hospital,
Level 4, The O'Brien Centre St. Vincent's Hospital, 394-404 Victoria Street Darlinghurst, Sydney, NSW, 2010, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 September 2013
Received in revised form
25 March 2014
Accepted 23 May 2014
Available online 2 June 2014
We explored whether transdiagnostic internet-delivered cognitive behavioural therapy (iCBT) for mixed
anxiety and depression effectively reduces repetitive negative thinking (RNT), and whether reductions in
RNT and positive metacognitive beliefs mediate symptom improvement during iCBT.
Participants with generalized anxiety disorder (GAD), major depressive disorder (MDD), or mixed
GAD/MDD diagnoses were randomly allocated to a 6-lesson clinician-guided iCBT anxiety and depression
program (n ¼ 46) or wait-list control (WLC, n ¼ 53). Depression (PHQ-9), generalized anxiety (GAD-7),
RNT (Repetitive Thinking Questionnaire) and positive beliefs about RNT (Positive Beliefs about Rumination Scale) were assessed at pre-, mid-, and post-treatment or matched time points for WLC. Tests of
serial indirect effects explored the potential meditating role of RNT and positive belief reductions on the
impact of iCBT on depression and anxiety symptoms post-treatment.
Results showed that both RNT frequency and positive beliefs about the value of RNT reduced significantly following iCBT compared to WLC, with gains maintained at 3-month follow-up. Reductions between pre- and mid-treatment in positive beliefs and RNT mediated improvements in depression
symptoms post-iCBT, and reductions in positive beliefs mediated improvements in GAD symptoms.
These ﬁndings indicate that iCBT is an effective treatment for RNT and positive metacognitive beliefs.
Future dismantling studies are needed to assess the most effective treatment components.
© 2014 Elsevier Ltd. All rights reserved.
Keywords:
Cognitive behavioural therapy
Depression
Anxiety
Treatment
Rumination
Worry
Introduction
Major depressive disorder (MDD) and generalized anxiety disorder (GAD) are disabling and highly co-morbid conditions (Hunt,
Slade, & Andrews, 2004) that contribute substantially to the
global burden of disease (World Health Organization, 2008). MDD
and GAD have some of the highest comorbidity rates of the internalising disorders cluster, reaching as high as 58e70% (Brown,
Campbell, Lehman, Grisham, & Mancill, 2001; Hunt et al., 2004).
Although they are distinct disorders, they are closely related
(Watson, 2005); MDD and GAD share similar genetic and environmental vulnerabilities (Hettema, 2008), are both characterized
by trait negative affectivity or neuroticism (Andrews, 1990), share
overlapping symptoms (e.g., sleep disturbance, APA, 1994), and
* Corresponding author. Tel.: þ61 2 8382 1433.
E-mail addresses: [email protected] (J.M. Newby), alishia.williams@unsw.
edu.au (A.D. Williams), [email protected] (G. Andrews).
http://dx.doi.org/10.1016/j.brat.2014.05.009
0005-7967/© 2014 Elsevier Ltd. All rights reserved.
similar cognitive and behavioural maintenance processes (e.g.,
avoidance).
One of the core cognitive processes that characterizes and perpetuates both disorders is repetitive negative thinking (RNT) in the
form of worry and rumination. Although there appear to be some
minor differences between rumination and worry in features such
as temporal orientation (i.e., past versus future focus) (NolenHoeksema, Wisco, & Lyubomirsky, 2008), there is growing evidence that rumination and worry are highly similar cognitive
processes (Watkins & Moulds, 2005) that are both associated with a
broad construct of negative repetitive thought (RNT). RNT,
capturing both rumination and worry, has been conceptualized as
perseverative, frequent, and uncontrollable cognitive activity
focused on negative aspects of the self and world (Ehring &
Watkins, 2008; Segerstrom, Stanton, Alden, & Shortridge, 2003).
RNT is a critical transdiagnostic maintaining factor across many
Axis I disorders, including social phobia and post-traumatic stress
disorder (PTSD), as well as GAD and MDD (Aldao & NolenHoeksema, 2010; Harvey, Watkins, Mansell, & Shafran, 2004;
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
Watkins, 2009). There is substantial evidence that RNT has negative
cognitive, emotional and behavioural consequences (e.g.,
Lyubomirsky & Nolen-Hoeksema, 1995), and predicts the onset and
maintenance of anxiety and depressive symptoms (NolenHoeksema, 2000). In addition, although the degree to which RNT
as a transdiagnostic construct predicts treatment response has not
been studied, the tendency to ruminate has been associated with a
poorer response to depression treatment (Schmaling, Dimidjian,
Katon, & Sullivan, 2002), slower rates of recovery, and greater
likelihood of premature drop-out during depression treatment
(Crane & Williams, 2010). Collectively this body of research indicates that maladaptive RNT, and the factors which drive RNT need
to be a core focus of treatment, particularly in treatments for comorbid depression and anxiety (McLaughlin & Nolen-Hoeksema,
2011).
There have been promising efforts to develop ruminationfocused treatments for depression (e.g., rumination-focused CBT
and concreteness training; Watkins et al., 2007, 2012), and worryfocused treatments for GAD (e.g., Metacognitive therapy for GAD,
Wells, 1995) that speciﬁcally target rumination and worry,
respectively. Cognitive Behavioural Therapy (CBT) has also been
shown to effectively reduce the frequency of rumination in
depression (Jones, Siegle, & Thase, 2008; Manicavasagar, Perich, &
Parker, 2012) and pathological worry in GAD (Covin, Ouimet, Seeds,
& Dozois, 2008). However, few augmentations to therapy have
been made to address maladaptive RNT processes from a transdiagnostic perspective. In addition, the degree to which transdiagnostic treatments for mixed depressed and anxious samples
effectively reduces RNT and the factors which drive the tendency
for RNT still needs to be evaluated.
Various contemporary theories of RNT (Dugas, Gagnon,
Ladouceur, & Freeston, 1998; Nolen-Hoeksema, 1991; Papageorgiou & Wells, 2001; Wells, 1995; Wells & Matthews, 1996) such
as the Self-Regulatory Executive Function model (S-REF) (Wells &
Matthews, 1994) propose that positive metacognitive beliefs
which emphasize the perceived advantages of RNT (e.g., “rumination helps me understand and solve my problems,” “worrying helps
me prevent future catastrophes”) reinforce the tendency to use RNT
as a coping strategy. In support for the hypothesized role of positive
metacognitive beliefs, research has demonstrated that positive
metacognitive beliefs about RNT are prevalent in GAD, MDD and
comorbid GAD/MDD (Dugas et al., 1998; Dupuy & Ladouceur, 2008;
Papageorgiou & Wells, 2001), are associated with increased frequency of RNT (as indexed by worry and rumination) (CartwrightHatton & Wells, 1997) and predict symptom maintenance
(Papageorgiou & Wells, 2001). Studies of patients with GAD have
also shown that directly challenging positive beliefs about the
utility, function and consequences of worry reduces worry frequency (e.g., Robinson et al., 2010), suggesting positive beliefs are
an important clinical target.
With these empirical ﬁndings in mind, we developed a transdiagnostic cognitive behavioural treatment for comorbid MDD and
GAD to be delivered over the internet (iCBT), which placed a focus
on reducing RNT and positive metacognitive beliefs. There are
many advantages to developing such a program: iCBT offers an
effective low-intensity mode of treatment that is easily disseminated and accessible to a large number of patients in the community (Andersson & Cuijpers, 2009; Andrews, Cuijpers, Craske,
McEvoy, & Titov, 2010; Spek et al., 2007). iCBT has been shown to
have similar efﬁcacy to face-to-face CBT (Andrews, Davies, & Titov,
2011), is cost-effective, and symptom improvements are maintained over long-term, even up to 5-year follow-up for some programs (Hedman et al., 2011). Transdiagnostic treatments also
present a novel approach to address comorbidity across the internalizing disorders. Such treatments aim to target shared symptoms
53
and maintaining factors across anxiety and depressive disorders
(e.g., Barlow et al., 2004), and have received growing support in
various treatment delivery modes including when delivered individually (Barlow, Allen, & Choate, 2004), in a group format (Norton,
2012), and via the internet (Carlbring et al., 2011; Titov et al., 2011).
Using a randomised controlled trial (RCT) design, we demonstrated
large effect size reductions in both depression and generalized
anxiety symptoms following our iCBT program compared to a waitlist control (see Newby et al., 2013). In the current study, using data
drawn from the RCT, we sought to test whether this iCBT program
was also effective in reducing RNT, and whether it performed better
than wait-list control in reducing RNT. We measured RNT frequency
at pre, mid, and post-treatment using the Repetitive Thinking
Questionnaire (RTQ) (McEvoy, Mahoney, & Moulds, 2010), a
symptom-independent measure of RNT that has been validated in
non-clinical and clinically anxious and depressed samples. We also
aimed to test whether iCBT reduced positive metacognitive beliefs
about RNT; one of the factors that are hypothesized to drive the
tendency to engage in maladaptive RNT in MDD and GAD.
In the current study, we also aimed to extend our understanding
beyond the efﬁcacy of our iCBT program, by exploring whether
reductions in positive metacognitive beliefs and RNT frequency
would mediate the impact of iCBT on symptoms of both depression
and generalized anxiety. We tested for serial indirect effects (Hayes,
2013) to investigate this possibility. Because the CBT skills that
aimed to target and alleviate negative repetitive thinking and
positive beliefs about the value of repetitive thinking were included
in the ﬁrst half of treatment (lesson 1e3), we placed our focus on
analysing whether changes in these variables in the ﬁrst half of
treatment mediated the impact of treatment on depression and
generalized anxiety symptom outcomes. We hypothesized that by
targeting positive beliefs about the value of RNT, patients in the
iCBT group would experience reductions in beliefs that would lead
to a reduction in RNT frequency, and in turn, lead to symptom
reduction. To our knowledge, no previous trials have investigated
the impact of CBT (in online or face-to-face modality) on a transdiagnostic measure of RNT and positive metacognitive beliefs, nor
investigated the mediating role of RNT from a transdiagnostic
perspective on symptom improvement.
In summary, the aims of this study were to examine whether: 1)
transdiagnostic iCBT for mixed anxiety and depression was more
effective than wait-list control in reducing RNT and positive metacognitive beliefs about RNT, and whether 2) reductions in RNT
frequency and positive beliefs about RNT mediate the impact of
iCBT on depression and generalized anxiety. We hypothesized that
iCBT would be more effective than wait-list control at effectively
reducing both RNT frequency and positive metacognitive belief
ratings. Second, we predicted that the effect of iCBT on primary
symptoms of depression and generalized anxiety would at least
partially be mediated by the reductions in RNT and positive metacognitive belief ratings.
Methods
Design
A CONSORT 2010 compliant (Schulz et al., 2010) RCT design
compared a group who received immediate iCBT (iCBT group) to a
wait-list control group (WLC). The iCBT group was followed up until
3 months post-treatment. The WLC group commenced the iCBT
program after the post-treatment assessment. The study was
approved by the Human Research Ethics Committee (HREC) of St
Vincent's Hospital (Sydney, Australia) (HREC 11/SVH/95), and the
trial was registered as ACTRN12611001055998. For more extensive
54
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
detail about the study protocol, design, attrition and excluded
participants, see Newby et al. (2013).
Participants and randomisation
Participants (>18 years, who self-identiﬁed as suffering from
GAD, MDD or mixed anxiety and depression) were recruited from
an existing wait-list for iCBT and online advertisement, and
completed online screening at www.virtualclinic.org.au. Online
screening included the PHQ-9, and GAD-7, as well as self-report
items that assessed the exclusion criteria. Exclusion criteria were:
self-reported history of a psychotic disorder, bipolar disorder,
substance or alcohol dependence, self-reported history of suicide
attempts or frequent suicidal ideation (on more than half of the
days in the past two weeks), severe depression symptoms (PHQ-9
score > 20), current benzodiazepine or atypical antipsychotic use.
Participants who were taking current antidepressant medication or
receiving current psychological treatment were not excluded, but
must have been on a stable dosage for at least 2 months prior to
participating, and agreed to not make any changes to their treatment during the entire duration of the study.
Participants who had GAD-7 and PHQ-9 scores above clinical
threshold (greater than or equal to 10), and met the study's inclusion criteria were then interviewed by telephone by trained interviewers to conﬁrm the presence of DSM-IV GAD and/or MDD
using the Mini International Neuropsychiatric Interview (MINI,
Version 5.0.0, Sheehan et al., 1998). Following online screening and
telephone interview administration of the MINI, eligible participants were randomised to iCBT (n ¼ 49) or WLC (n ¼ 60).1
Description of treatment
The iCBT program is a clinician-assisted treatment delivered via
www.virtualclinic.org.au, which consists of six illustrated online
lessons and a corresponding lesson summary (which includes
practical homework exercises), and is completed over a 10-week
period. Lesson content is presented in the form of an illustrated
story about two ﬁctional characters who experience anxiety and
depression, and gain mastery over their symptoms with the help of
a clinician using CBT techniques. The speciﬁc components that
were designed to target RNT and positive beliefs were: psychoeducation, self-monitoring, identiﬁcation and cognitive restructuring of positive and negative beliefs about RNT, behavioural
activation, structured problem-solving, attention shifting, worry
time, and imaginal exposure to worry scenarios. See Table 1 for
lesson-by-lesson content. All patients received the same standardized treatment. Following each lesson, participants download
and print out a lesson summary, which includes practical homework exercises to complete before the next lesson (e.g., graded
exposure tasks). Participants also have access to frequently asked
questions, former patients' recovery stories, and extra resources on
a range of topics such as sleep hygiene, medications and panic attacks. Once participants downloaded their homework, they were
considered to have ‘completed the lesson.’ A new lesson was made
available each week for the ﬁrst six weeks to encourage completion
of one lesson before moving onto the next lesson. Participants were
encouraged to practice their lesson homework for at least one week
1
To randomly allocate participants to the relevant group, 110 random numbers (1
or 2 to determine group allocation) were generated using www.random.org (a tool
which does not necessarily restrict group allocation to ensure there are even
numbers across groups) by a research assistant independent from the trial. Group
allocation was concealed in an opaque envelope, which was only opened at the
point when an offer of treatment was made.
Table 1
Lesson content of the internet-delivered cognitive behavioural treatment program.
Lesson
number
Content
Homework tasks
1
Psychoeducation about anxiety and
depression, the ﬁght or ﬂight response,
controlled breathing, and physical exercise
Cognitive therapy components: education
about the cognitive model, cognitive
distortions, and introduction to thought
monitoring; activity planning
Thought challenging/cognitive
restructuring; challenging positive and
negative metacognitive beliefs about
repetitive negative thinking; shifting
attention, hunt for positives
Education about avoidance and safety
behaviours; graded exposure and
structured problem solving
Advanced graded exposure (imaginal
exposure, interoceptive exposure);
troubleshooting difﬁculties with graded
exposure
Relapse prevention
Controlled breathing,
physical exercise
2
3
4
5
6
Thought monitoring,
activity planning
Thought challenging,
hunt for positives
Graded exposure and
structured problem
solving
Graded exposure
Relapse prevention
plan
(and two weeks each for lessons 4 and 5 which both included
exposure tasks) before starting the next lesson. Participant
comprehension of the lesson content was not assessed.
Clinician guidance
The primary clinician was the practice manager of CRUfAD (an
administrative assistant), who was supervised by a Therapist (JN, a
PhD-level Clinical Psychologist). The iCBT group participants
received regular email and/or phone contact with their Clinician
until they completed lesson 2, after which contact was made in
response to patient request. All emails requiring clinical advice
were responded to by the Therapist. If clinically indicated, or if
patients' Kessler-10 (K-10) and/or PHQ-9 scores deteriorated
(increased by at least 0.5 SD), the Therapist would make telephone
contact with the participant. The Kessler-10 (Kessler et al., 2002)
was administered prior to each lesson to keep track of lesson-bylesson progress and enable early detection of any increase in
symptoms, but was not a primary outcome measure in the current
study.
Outcome measurement
Outcome measures were administered at pre-, mid-treatment
(prior to commencing lesson 4), post-treatment, and 3-month
follow-up (for the iCBT group only). The WLC participants
commenced iCBT immediately after post-treatment assessment,
and thus did not provide 3-month follow-up data. Forty six iCBT
and 53 WLC participants completed pre-treatment questionnaires
and were included in analyses. All of the WLC participants provided
complete post-treatment data. Forty three out of 46 and 40/46 had
complete post-treatment and 3-month follow-up data respectively
in the iCBT group.
Outcome measures included the Patient Health Questionnaire-9
(depression) (PHQ-9, Kroenke, Spitzer, & Williams, 2001) and the
Generalized Anxiety Disorder 7-item scale (generalized anxiety)
(GAD-7, Spitzer, Kroenke, Williams, & Lowe, 2006). The 10-item
version of the Repetitive Thinking Questionnaire (McEvoy et al.,
2010) assessed RNT frequency. This measure has been validated
in clinical samples and has good internal reliability (a ¼ .89)
(Mahoney, McEvoy, & Moulds, 2012). Items (e.g., ‘Once I started
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
thinking about the situation, I couldn't stop’) were rated on a 5point scale about how they tended to respond to distressing situations (1 ¼ not true, to 5 ¼ very true). The 9-item Positive Beliefs
about Rumination Scale-Adapted Version (PBRS-A; Watkins &
Moulds, 2005) measured positive beliefs about RNT more generally, as opposed to ruminating about depression which was
assessed with the original version of the PBRS (Papageorgiou &
Wells, 2001). Participants' agreement with each item (e.g., ‘I
need to think about things to ﬁnd answers to how I feel’) was rated
on a 4-point scale generally (1 ¼ do not agree to 4 ¼ agree very
much). This measure has good internal reliability (a ¼ .89;
Watkins & Moulds, 2005).
Statistical analyses
Baseline comparisons
Independent samples t-tests and chi-square (for nominal data)
were used to compare the groups at pre-treatment to examine
whether there were any baseline group differences.
Intent-to-treat (ITT) marginal model analyses
To examine the impact of iCBT on depression and anxiety
symptoms, as well as the frequency of RNT and positive beliefs, we
conducted separate analyses for each outcome measure (the PHQ9, GAD-7, RTQ, PBRS-A ) with group as a ﬁxed factor and time as a
repeated factor. Marginal model analyses were used to account for
missing data due to participant drop-outs. Marginal models are
appropriate for pre-post designs and RCTs with multiple time
points (Salim, Mackinnon, Christensen, & Grifﬁths, 2008). This
analysis has added advantages over ITT analyses using the last
observation carried forward, because it does not assume that the
last measurement was stable (Gueorguieva & Krystal, 2004). As the
primary outcome measures were also collected at mid-point, effects for the primary measures were modelled using the restricted
maximum likelihood (REML) model estimation method with an
autoregressive (AR1) covariance structure to account for the correlation between the time-points. Signiﬁcant effects were followed
up with pairwise contrasts comparing pre-treatment to posttreatment scores.
Mediation analyses
Tests of the indirect effects (mediation) were conducted using
PROCESS (Hayes, 2013). This method was chosen over the causal
steps approach (Baron & Kenny, 1986) based on recent research
advocating for the use of modern statistical approaches to
quantifying intervening variable models (Hayes, 2009). As recommended, particularly for small samples, estimates of indirect
effects were generated using bootstrapping analysis (see Preacher
& Hayes, 2004; Williams & MacKinnon, 2008). Bootstrapping is a
nonparametric resampling method that generates an estimate of
the indirect effect, and does not require assumptions about the
shape of the sampling distribution that underlie the Sobel test. In
bootstrapping analysis, the most stringent test of an indirect effect (mediation) is if the 95% bias corrected and accelerated
conﬁdence intervals for the indirect effect do not include the
value of 0. When zero is outside of the 95% conﬁdence interval
estimate, the indirect effect is declared statistically different from
zero at p < .05 (two-tailed), indicating that the effect of the independent variable on the dependent variable is contingent upon
the effect of the proposed mediator (Hayes & Preacher, 2010;
Preacher & Hayes, 2004). In the current study, we estimated 5000
bias-corrected bootstrap 95% conﬁdence intervals using PROCESS
for SPSS.
55
Results
Baseline
Participants were on average 44 years old (SD ¼ 12.2) and
moderately depressed and anxious (see Table 2). The majority were
female (77.8%), married or de-facto (62.6%), were educated with a
bachelor or postgraduate degree (58.6%), in full-time (36.4%) or
part-time paid work (23.2%). Groups were matched on demographics, scores on outcome measures (see Table 2) and diagnostic status (ps > .05). The majority had co-morbid GAD/MDD
(47%) followed by GAD with sub-threshold MDD (38%). As determined by the MINI interviews, in the iCBT group, 19 met criteria for
co-morbid GAD/MDD (41.3%), 21 had GAD (45.7%) (but with subthreshold MDD), and 6 met criteria for MDD (13.0%) and had subthreshold GAD. In the WLC group, 28 met criteria for co-morbid
GAD/MDD (52.8%), 16 (30.2%) had GAD, and 9 (17.0%) had MDD,
but each had sub-threshold GAD or MDD respectively. There were
no differences in the proportions of participants per group who had
a diagnosis of GAD, MDD or co-morbid GAD/MDD (c2(2) ¼ 2.52,
p > .05). We deﬁned sub-threshold diagnoses of GAD and MDD
when a participant had an elevated score (>5) on the relevant selfreport measure (GAD-7, and PHQ-9), but did not endorse sufﬁcient
diagnostic criteria on the MINI to warrant a diagnosis of GAD or
MDD.
Adherence, outcome measures and effect sizes
Of the 46 participants in the iCBT group, 41 completed the total
six lessons, representing an adherence rate of 89%. The main effects
for all outcome measures were qualiﬁed by signiﬁcant
Group # Time interactions, p's $ .001 (see Table 2 for observed
means, standard deviations, statistics and effect sizes). Betweengroup comparisons on all measures revealed that post-treatment
scores were signiﬁcantly lower in the iCBT group relative to WLC,
with moderate to large observed effect sizes (Cohen's d) (ranging
from .78 (PBRS-A) to 1.00 (PHQ-9)). Within-group comparisons
between pre- and post-treatment for the iCBT group revealed large
effect sizes (Hedges g) (range ¼ .96 (GAD-7) to 1.40 (RTQ)). The
within-group reductions in the WLC group on all measures were
not signiﬁcant (ps > .05).
Clinically signiﬁcant change
Following Jacobson and Truax (1991), we calculated reliable
change index values for the RTQ and PBRS-A, and found reliable
improvement at post-treatment in greater proportions of the iCBT
group than the WLC group using chi square analyses (ps < .001). For
the iCBT group, 55% evidenced reliable change on the RTQ and 35%
on the PBRS-A; whereas in the WLC group only 7% showed reliable
change on the RTQ and on the PBRS-A. Seven percent of WLC
participants indexed reliably increased RTQ and PBRS-A at posttreatment.
Three-month follow-up data (iCBT group only)
For the iCBT group, marginal model analyses with Time as variable ﬁxed factor and baseline scores entered as a covariate were
conducted for each DV (PHQ-9, GAD-7, RTQ AND PBRS-A) to
compare mean score reductions between post-treatment to 3month follow-up. For GAD-7 and RTQ scores, the main effects of
Time were signiﬁcant, corresponding to a small effect size for reductions in GAD-7 scores (ES ¼ .26, F (1,37.36) ¼ 4.39, p ¼ .04,
r ¼ .67) and a moderate effect size for RTQ scores (ES ¼ .48, F
(1,35.67) ¼ 13.30 p ¼ .001, r ¼ .64). Follow-up RTQ scores in the iCBT
56
5.85
17.62
17.90
5.97
17.66
6.13
e
e
t(97) ¼ .51,
p ¼ .60
7.43
12.09
5.88
12.87
6.73
5.84
18.17
15.95
8.89
37.33
37.54
8.30
35.94
8.63
e
e
t(97) ¼ .23,
p ¼ .81
10.11
21.52
9.86
27.60
8.67
7.56
37.82
34.72
5.00
10.43
9.30
4.77
9.92
4.90
e
e
t(178.02) ¼ .06,
p ¼ .94
3.71
4.39
4.28
5.93
4.63
3.74
10.37
8.11
e
e
4.88
10.41
5.52
11.24
4.80
11.62
Note. PHQ-9 ¼ The Patient Health Questionnaire-9, GAD-7 ¼ Generalized Anxiety Disorder 7-item scale, RTQ ¼ Repetitive Thinking Questionnaire, 10-item scale, PBRS-A ¼ Positive Beliefs about Rumination Scale-Adapted
Version, iCBT ¼ internet-based cognitive behavioural treatment group, WLC ¼ wait-list control group, M ¼ mean, SD ¼ standard deviation, Pre ¼ pre-treatment scores, Mid ¼ mid-treatment scores prior to lesson 4,
post ¼ post-treatment scores, 3-months ¼ scores at 3-month follow-up.
e
.005 (.37e.38)
.55
.78 (.36e1.21)
.96 (.52e1.39)
F(1, 95.51) ¼ 14.60,
p < .001
.46
e
.13 (.25e.51)
.71
.90 (.47e1.33)
1.40 (.95e1.84)
F(1, 95.11) ¼ 19.63,
p < .001
.39
e
.07 (.31e.45)
.70
.85 (.43e1.27)
.96 (.53e1.38)
F(1, 187.16) ¼ 19.36,
p < .001
.44
e
.26 (.12e.64)
.61
1.00 (.59e1.40)
F(1, 166.32) ¼ 26.51,
p < .001
t(229.84) ¼ 7.05,
p < .001
t(228.66) ¼ 1.94,
p ¼ .15
t(240.55) ¼ 5.95,
p ¼ .001
t(239.31) ¼ .72,
p ¼ .99
t(94.00) ¼ 8.03,
p < .001
t(91.41) ¼ 1.25,
p ¼ .21
t(96.28) ¼ 5.87,
p < .001
t(92.73) ¼ .04,
p ¼ .96
t(158.34) ¼ 1.31,
p ¼ .18
3.79
4.05
4.24
5.76
4.94
7.93
3.90
10.39
PHQ-9
iCBT
PHQ-9
WLC
GAD-7
iCBT
GAD-7
WLC
RTQ
iCBT
RTQ
WLC
PBRS-A
iCBT
PBRS-A
WLC
SD
Measure
M
M
SD
M
SD
M
SD
Pre-treatment
comparisons
Within-group comparison
(pre to post)
Between-group
comparison (post)
.52
1.05 (.62e1.47)
Between-group
(pre to post)
Within-group
(pre to post)
Effect size
Hedges g (95%CI)
Effect size
Cohen's d (95%CI)
r (for withingroup ES)
F (df)
t (df)
t (df)
3-months
Post
Mid
Pre
Table 2
Observed means, standard deviations and effect sizes for depression, anxiety, negative repetitive thinking and positive beliefs about repetitive thinking before and after internet-based cognitive behavioural treatment.
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
group (M ¼ 21) are lower than a mean of 34.03 reported in a clinical
sample (Mahoney et al., 2012) and a mean of 27.06 in an unselected
student sample (McEvoy et al., 2010). For 3-month follow-up PHQ9 scores and PBRS-A scores, the main effects of Time were not
signiﬁcant (PHQ-9: F (1,36.56) ¼ 3.04, p ¼ .09, r ¼ .64; PBRS-A: F
(1,38.39) ¼ .80, p ¼ .37, r ¼ .64).
Indirect effects (serial multiple mediation analyses)
The mediating role of reductions in positive beliefs and repetitive
thinking between pre- and mid-treatment on post-treatment
outcome (depression and generalized anxiety)
Depression symptoms. First, we examined whether the reductions
in positive metacognitive beliefs and RNT between pre- and midtreatment mediated the impact of iCBT on post-treatment
depression symptoms. To explore this question, in the ﬁrst
model, we estimated a serial multiple mediator model with two
theoretically-derived mediators [M1: changes in positive metacognitive belief ratings (PBRS-A scores) and M2: change in RNT
frequency (RTQ scores) between pre-treatment and midtreatment] using PROCESS (Hayes, 2013). This enabled us to test
the serial indirect effects of treatment on post-treatment depression (PHQ-9) scores via changes in positive metacognitive belief
ratings and change in RNT scores (see Fig. 1 for model). Because
reductions in positive beliefs were hypothesized to contribute to
reductions in RNT, positive belief change scores were entered as
the ﬁrst mediator in the proposed causal chain. Baseline PHQ-9
scores were entered as a covariate. A negative change score between pre- and mid-treatment indicates a reduction in positive
belief ratings (or RNT scores). The treatment variable (X) was
coded as 0 ¼ wait-list control, and 1 ¼ iCBT. All 95% bias-corrected
bootstrap conﬁdence intervals for the speciﬁc indirect effects are
presented in brackets.
Results indicated that the total effect of treatment on posttreatment PHQ-9 scores was signiﬁcant (Effect ¼ %4.55,
t ¼ %5.96, SE ¼ .76, p < .001). The predicted model of the serial
indirect effect of treatment on post-treatment depression via reductions in positive beliefs and reductions in RNT scores (between
pre- and mid-treatment) was supported, as the indirect effect was
statistically different from zero (95%CI: %.22 to %.001). The indirect
effect of treatment on depression via change in positive beliefs
(between pre- and mid-treatment) was also statistically different
from zero (95%CI ¼ %.88 to %.04), but the indirect effect of treatment on depression via changes in RNT scores was not (95%CI: %.73
to .03). When we repeated the same analysis with gender (coded
female ¼ 0, and male ¼ 1) and baseline scores on the PHQ-9, RTQ
and PBRS-A entered as covariates, the serial indirect effect of
treatment on post-treatment depression scores via reductions in
positive beliefs and repetitive thinking scores (between pre- and
mid-treatment) remained signiﬁcant (95%CI: %.42 to %.02), but the
indirect effect of treatment via reductions in positive beliefs was no
longer signiﬁcant (95%CI: %.60 to .13).2
These results suggest that the early reductions in positive
metacognitive belief ratings mediated the positive impact of iCBT
on post-treatment depression symptoms by inﬂuencing reductions
in RNT frequency. However, the two mediators were examined at
the same time-point, meaning we would be unable to rule out the
alternative causal pathway; reductions in RNT inﬂuencing reductions in positive beliefs. To test this model, we repeated the
2
Within this analysis, when all other variables were held constant, out of the
covariates included in the model, only baseline PHQ-9 scores signiﬁcantly predicted
post-treatment PHQ-9 scores (p < .001), with higher baseline severity predicting
higher severity at post-treatment.
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
57
Fig. 1. Serial multiple mediation model with post-treatment PHQ-9 scores as the
outcome variable, pre-treatment PHQ-9 scores entered as a covariate, and change in
positive beliefs entered as the ﬁrst mediator. Numbers represent beta values. Treatment is coded 0 ¼ wait-list control, 1 ¼ internet-delivered cognitive behavioural
therapy. DPos Beliefs ¼ changes between pre-treatment and mid-treatment (prior to
lesson 4) on the Positive Beliefs about Rumination Scale (adapted version); lower
scores indicate greater reductions in beliefs. DRNT ¼ changes on the Repetitive
Thinking Questionnaire, where lower scores indicate greater reductions in repetitive
thinking frequency. PHQ-9 (post) ¼ post-treatment scores on the Patient Health
Questionnaire 9-item scale.* ¼ signiﬁcant pathway; alpha level < .05.
Fig. 2. Serial multiple mediation model with post-treatment GAD-7 scores as the
outcome variable, pre-treatment GAD-7 scores entered as a covariate, and change in
positive beliefs entered as the ﬁrst mediator. Numbers represent beta values. Treatment is coded 0 ¼ wait-list control, 1 ¼ internet-delivered cognitive behavioural
therapy. DPos Beliefs ¼ changes between pre-treatment and mid-treatment (prior to
lesson 4) on the Positive Beliefs about Rumination Scale (adapted version); lower
scores indicate greater reductions in beliefs. DRNT ¼ changes on the Repetitive
Thinking Questionnaire, where lower scores indicate greater reductions in repetitive
thinking frequency. PHQ-9 (post) ¼ post-treatment scores on the Patient Health
Questionnaire 9-item scale. * ¼ signiﬁcant pathway; alpha level < .05.
above analyses with changes in RNT entered as the ﬁrst mediator.
The serial indirect effect of interest (treatment on depression scores
via change in RNT and change in positive beliefs) was not statistically different from zero (95%CI: %.18 to .001). We found the same
results when gender and baseline scores on the PHQ-9, RNT, and
positive belief ratings were entered as covariates (95%CI: %.26
to .01).
The mediating role of reductions in positive beliefs and repetitive
thinking between pre- and post-treatment on post-treatment
outcomes (depression and generalized anxiety)
Next, we repeated the same analyses to examine whether the
reductions in RNT and positive beliefs throughout the entire program (between pre- and post-treatment) mediated the impact of
iCBT on post-treatment outcomes on the PHQ-9 and GAD-7.
Although we cannot establish a causal relationship between the
proposed mediating variables and outcome variables, this analysis
provides a preliminary step towards understanding how reductions
in repetitive thinking and positive beliefs throughout the entire
course of treatment inﬂuenced outcomes. In this analysis, we
entered gender, and baseline PHQ-9, RNT and positive belief scores
as covariates. Positive belief change scores (from pre- to posttreatment) were entered as the ﬁrst mediator, and RNT change
scores (from pre- to post-treatment) as the second mediator in the
proposed causal chain. In this model, the serial indirect effect of
treatment on post-treatment depression via reductions in positive
beliefs and reductions in RNT scores was supported, as the indirect
effect was statistically different from zero (95%CI: %.72 to %.01).
The indirect effect of treatment on depression via change in RNT
scores was also statistically different from zero (95%CI ¼ %2.49
to %.58). However, the indirect effect of treatment on depression
via change in positive belief scores was not statistically different
from zero (95%CI ¼ %.58 to .98). These results suggest that when
gender, and baseline symptom severity, repetitive thinking frequency and positive beliefs ratings are controlled for, the positive
effects of iCBT on post-treatment depression symptoms is mediated
in part by reductions in positive beliefs inﬂuencing reductions in
repetitive thinking, and in part by direct reductions in repetitive
thinking4
We repeated the same serial multiple mediation analysis with
post-treatment GAD-7 scores as the outcome variable, and with
gender, baseline GAD-7 scores, baseline RNT scores and baseline
positive belief scores as the covariates. Again, changes in positive
beliefs were entered as the ﬁrst mediator, and RNT change scores as
the second mediator variable in the causal chain. Results showed
that the total effect of treatment on GAD-7 scores was signiﬁcant
Generalized anxiety symptoms. Next, we sought to examine
whether the reductions in positive metacognitive beliefs and RNT
between pre- and mid-treatment mediated the impact of iCBT on
post-treatment generalized anxiety symptoms. To examine this
question we repeated the same serial multiple mediation analysis
with post-treatment GAD-7 scores as the outcome variable, pretreatment GAD-7 scores entered as a covariate, with positive
belief change scores (from pre- to mid-treatment) entered as the
ﬁrst mediator, and RNT change scores (from pre- to midtreatment) as the second mediator in the proposed causal chain
(see Fig. 2). The total effect of treatment on GAD-7 scores was
signiﬁcant (Effect ¼ %4.54, t ¼ %5.91, SE ¼ .77, p < .001). The
predicted model of the serial indirect effect of treatment on GAD-7
symptoms via reductions in positive beliefs and reductions in RNT
(between pre- to mid-treatment) was not different from zero (95%
CI ¼ %.17 to .01), nor was the indirect effect of treatment via RNT
scores (95%CI ¼ %.53 to .07). However, the indirect effect of
treatment on post-treatment GAD-7 scores via change in positive
belief scores was statistically different from zero (95%CI ¼ %1.23
to %.05). These results suggest that iCBT had a positive impact on
generalized anxiety symptoms through reducing positive beliefs
about the value of RNT between pre- and mid-treatment. When
we carried out the same analysis with gender, baseline GAD-7,
baseline RNT and baseline positive belief scores as the covariates, none of the indirect effects were statistically different from
zero. Results were as follows: the serial indirect effect via change
in positive beliefs and change in RNT scores: (95%CI ¼ %.32 to .01),
the indirect effects via change in positive belief scores: (95%
CI ¼ %1.06 to .14), and the indirect effect via change in RNT scores
(95%CI ¼ %.72 to .04).3
3
Within this analysis, baseline positive belief scores and baseline GAD-7 scores
were signiﬁcant predictors of post-treatment GAD-7 scores (ps < .05), when all
other variables were held constant, such that higher scores predicted greater reductions in symptoms.
4
Within this analysis, when all other variables were held constant, both baseline
RNT scores and baseline PHQ-9 scores signiﬁcantly predicted post-treatment PHQ-9
scores (ps < .05), with higher scores predicting higher post-treatment PHQ-9
scores.
58
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
(Effect ¼ %4.35, t ¼ %5.94, SE ¼ .73, p < .001). The serial indirect
effect of treatment on GAD-7 symptoms via reductions in positive
beliefs and reductions in RNT scores between pre- and posttreatment was statistically different from zero (95%CI ¼ %.73
to %.02), as was the indirect effect of treatment via change in RNT
scores (95%CI ¼ %2.41 to %.49). However, the indirect effect of
treatment via changes in positive belief scores was not statistically
different from zero (95%CI ¼ %.71 to .85).5 These results for GAD
symptoms are similar to those for PHQ-9 symptoms, and suggest
that iCBT had a positive impact on generalized anxiety symptoms
through reducing positive beliefs about the value of repetitive
thinking which may in turn reduce the frequency repetitive
thinking throughout the program between pre- and posttreatment. Notably, within all of the mediation analyses conducted, we found no evidence of an effect of gender on post-treatment
outcome (depression and generalized anxiety) after controlling for
baseline symptoms severity and ratings of positive beliefs and RNT
frequency.
Discussion
We assessed the impact of a transdiagnostic internet-delivered
CBT program for mixed GAD and MDD on negative repetitive
thinking (RNT) and positive metacognitive beliefs about RNT in a
RCT comparing iCBT to a wait-list control condition. As hypothesized, we found signiﬁcant reductions in RNT frequency and ratings
of positive metacognitive beliefs about the value of RNT following
iCBT, with large observed effect sizes. We also found evidence for
further (albeit small) reductions in RNT between post-treatment
and 3-month follow-up for the iCBT group. Importantly, these results show clinician-assisted transdiagnostic iCBT program effectively reduces RNT and positive metacognitive beliefs, which are
two critically important maintaining factors in GAD and MDD
(Nolen-Hoeksema, 2000). We are unable to conclude which speciﬁc
components of the iCBT program were most effective in reducing
RNT and positive beliefs because of its' integral design (where patients received the same ﬁxed treatment). Dismantling studies are
now needed to isolate the key therapeutic ingredients so that
outcomes can be improved, particularly when considering that
45e60% of patients did not show clinically reliable change on RNT
and positive belief ratings at post-treatment.
We hypothesized that reductions in positive beliefs about RNT
and the frequency of RNT between pre-treatment and midtreatment would mediate the impact of treatment on therapeutic
outcomes (depression and anxiety symptoms) at post-treatment.
We chose to focus our examination on changes between pre- and
mid-treatment, rather than changes between pre- and posttreatment to establish temporal precedence between the mediators and outcome variables of interest (post-treatment depression
and anxiety scores), which is one of the criteria needed to establish
causality (Hayes, 2012). Notably, all of the core skills that specifically targeted RNT were included in the ﬁrst three lessons prior to
mid-treatment assessment, and were thus hypothesised to
contribute to the greatest change in these cognitive variables. Our
results were consistent with predictions for depression symptoms,
but not for generalized anxiety symptoms. That is, the results
suggested that relative to those in the WLC group, participants in
the iCBT group showed greater reductions in depression symptoms
at post-treatment, and these improvements were inﬂuenced by
greater reductions in positive beliefs observed at mid-treatment,
5
Within this analysis, baseline positive belief scores and baseline GAD-7 scores
were signiﬁcant predictors of post-treatment GAD-7 scores (ps < .05), such that
higher baseline levels predicted lower GAD symptoms after treatment.
which were associated with reductions in RNT, which in turn was
associated with improved depression outcomes.
We also tested the alternative causal model (that reduced RNT
frequency inﬂuenced the extent to which participants endorsed
positive beliefs about the value of RNT), and found no evidence for
this model. This preliminary analysis suggests that reductions in
positive beliefs inﬂuenced reductions in RNT and not vice versa,
although we acknowledge we cannot deﬁnitively rule out the
possibility of an alternative causal pathway between the two
mediating variables. Further research is needed to establish the
causal relationship between belief change and reductions in RNT
frequency during treatment, as well as the mechanism by which
these reductions promoted improvements in depression symptoms. Reduced RNT may have had a positive impact on a range of
different cognitive, affective and behavioural symptoms of
depression. For example, we would expect that reduced RNT about
the self, world and future would have a direct inﬂuence on
improving day-to-day mood, but also may reduce the tendency to
engage in maladaptive withdrawal and avoidance. In addition, it is
possible that the skills targeting RNT and positive beliefs were still
being consolidated after the mid-treatment time-point (and thus
leading to further improvement). The average scores on our measures of RNT and positive beliefs were lower at post-treatment than
mid-treatment, which is consistent with this possibility. Notably,
we found the same pattern of ﬁndings when we investigated
whether reductions between pre- and post-treatment in positive
beliefs and RNT mediated post-treatment depression symptoms.
However, the latter analysis provides a weaker test of the casual
relationship between these variables because the mediator and
outcome variables in were measured at the same time-point.
Future RCTs of iCBT which aim to explore mediation would
beneﬁt from inclusion of additional follow-up assessments to
examine the mediating role of these cognitive processes between
pre- and post-treatment on follow-up outcomes.
Unexpectedly, we found a different pattern of results for
generalized anxiety symptoms. Reductions in positive beliefs at
mid-treatment mediated the impact of treatment on generalized
anxiety symptoms, but not via reductions in RNT. This indirect effect was not observed when participants' baseline levels of RNT and
degree of positive beliefs were controlled for. It is unclear why the
results differed for depression versus generalized anxiety symptoms. We chose to conduct a more stringent test of mediation by
entering pre-treatment GAD-7 scores as a covariate in our analysis.
Because worry frequency forms a core feature of GAD and is
therefore assessed using the GAD-7, controlling for it in the mediation analysis may have obscured a potential mediating role of RNT
frequency. Alternatively, it is possible that it may take longer for a
shift in beliefs to lead to measurable changes in how frequently
patients engage in RNT for patients with primary GAD. These effects may have been detectable at the time of the 3-month followup, but unfortunately we could not test this proposal as the waitlist group had commenced iCBT at this time point. In addition, we
used the adapted version of the Positive Beliefs about Rumination
Scale, which has been adapted to assess positive beliefs about RNT
independent of depression symptoms. It was initially designed to
assess positive beliefs about rumination in depression, and may not
have captured critically important positive beliefs about RNT that
are tied to the maintenance of generalized anxiety symptoms,
which may be better captured by other existing measures such as
the Metacognitions Questionnaire which place a greater focus on
worry.
Our ﬁndings provide some additional support for cognitive
theories of MDD and GAD which argue for an important relationship between positive beliefs about RNT and RNT in the maintenance of depression and anxiety symptoms (e.g., Wells, 1995). We
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
do acknowledge that our study only explored two possible mechanisms by which iCBT inﬂuenced depression and anxiety outcomes.
Our study did not account for all possible mechanisms that may
have mediated symptom improvement. For example, there are a
large range of other cognitive, emotional and behavioural variables
that may have accounted for the reductions in depression and
generalized anxiety symptoms we observed in the iCBT group. For
instance, behavioural activation is an important clinical target and
may mediate the improvements during iCBT. In addition, it also
remains untested whether iCBT effectively reduces negative metacognitive beliefs about RNT which are also theorised to maintain
worry and rumination. Negative metacognitive beliefs emphasize
the perceived danger and harm related to engaging in RNT (e.g.,
‘this worrying is sending me crazy’). The potential role of these
variables in treatment outcomes awaits empirical investigation.
Future studies are also needed to explore the degree to which
the Repetitive Thinking Questionnaire, a relatively new transdiagnostic measure of RNT (McEvoy et al., 2010) captures
constructive versus unconstructive forms of RNT such as concrete/
experiential and abstract thinking (Watkins, 2008). Additionally,
the impact of our iCBT program on symptoms of other important
comorbid disorders, such as panic disorder and social anxiety disorder, await further examination. Notwithstanding these limitations, the current study provides evidence that iCBT for mixed GAD
and MDD is an effective treatment for RNT and positive metacognitive beliefs about RNT. Moreover, to our knowledge, our results provide the ﬁrst evidence that reductions in positive beliefs
about RNT, and for depression symptoms, reductions in RNT frequency are important mediators of the positive impact of iCBT on
therapeutic outcomes.
Acknowledgements
Jill M. Newby was supported by an Australian National Health
and Medical Research Council (NHMRC) Fellowship (1037787).
Alishia D. Williams was supported by an Australian National Health
and Medical Research Council (NHMRC) Fellowship (630746).
References
Aldao, A., & Nolen-Hoeksema, S. (2010). Speciﬁcity of cognitive emotion regulation
strategies: a transdiagnostic examination. Behaviour Research and Therapy,
48(10), 974e983. http://dx.doi.org/10.1016/j.brat.2010.06.002.
American Psychiatric Association (APA). (1994). Diagnostic and statistical manual of
mental disorders (4th ed.). Washington DC: APA.
Andersson, G., & Cuijpers, P. (2009). Internet-based and other computerized psychological treatments for adult depression: a meta-analysis. Cognitive Behaviour
Therapy, 38(4), 196e205. http://dx.doi.org/10.1080/16506070903318960.
Andrews, G. (1990). Classiﬁcation of neurotic disorders. Journal of the Royal Society
of Medicine, 83, 606e607.
Andrews, G., Cuijpers, P., Craske, M. G., McEvoy, P., & Titov, N. (2010). Computer
therapy for the anxiety and depressive disorders is effective, acceptable and
practical health care: a meta-analysis. PLoS ONE, 5(10), e13196. http://
dx.doi.org/10.1371/journal.pone.0013196.
Andrews, G., Davies, M., & Titov, N. (2011). Effectiveness randomized controlled trial
of face to face versus internet cognitive behaviour therapy for social phobia.
Australian and New Zealand Journal of Psychiatry, 45(4), 337e340. http://
dx.doi.org/10.3109/00048674.2010.538840.
Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a uniﬁed treatment for
emotional disorders. Behavior Therapy, 35(2), 205e230.
Baron, M., & Kenny, D. (1986). The moderatoremediator variable distinction in
social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173e1182. http://
dx.doi.org/10.1037/0022-3514.51.6.1173.
Brown, T. A., Campbell, L. A., Lehman, C. L., Grisham, J. R., & Mancill, R. B. (2001).
Current and lifetime comorbidity of the DSM-IV anxiety and mood disorders in
a large clinical sample. Journal of Abnormal Psychology, 110(4), 585e599. http://
dx.doi.org/10.1037/0021-843x.110.4.585.
Carlbring, P., Maurin, L., Torngren, C., Linna, E., Eriksson, T., Sparthan, E., et al. (2011).
Individually-tailored, internet-based treatment for anxiety disorders: a randomized controlled trial. Behaviour Research and Therapy, 49(1), 18e24. http://
dx.doi.org/10.1016/j.brat.2010.10.002.
59
Cartwright-Hatton, S., & Wells, A. (1997). Beliefs about worry and intrusions: the
meta-cognitions questionnaire and its correlates. Journal of Anxiety Disorders,
11(3), 279e296.
Covin, R., Ouimet, A. J., Seeds, P. M., & Dozois, D. J. (2008). A meta-analysis of CBT for
pathological worry among clients with gad. Journal of Anxiety Disorders, 22(1),
108e116. http://dx.doi.org/10.1016/j.janxdis.2007.01.002.
Crane, C., & Williams, J. M. (2010). Factors associated with attrition from
mindfulness-based cognitive therapy in patients with a history of suicidal
depression. Mindfulness, 1(1), 10e20. http://dx.doi.org/10.1007/s12671-0100003-8.
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(2), 215e226. http://dx.doi.org/10.1016/S0005-7967(97)00070-3.
Dupuy, J. B., & Ladouceur, R. (2008). Cognitive processes of generalized anxiety
disorder in comorbid generalized anxiety disorder and major depressive disorder. Journal of Anxiety Disorders, 22(3), 505e514.
Ehring, T., & Watkins, E. R. (2008). Repetitive negative thinking as a transdiagnostic
process. International Journal of Cognitive Therapy, 1(3), 192e205. http://
dx.doi.org/10.1521/ijct.2008.1.3.192.
Gueorguieva, R., & Krystal, J. (2004). Move over ANOVA: progress in analyzing
repeated-measures data and its reﬂection in papers published in the archives of
general psychiatry. Archives of General Psychiatry, 61(3), 310e317. http://
dx.doi.org/10.1001/archpsyc.61.3.310.
Harvey, A. G., Watkins, E., Mansell, W., & Shafran, R. (Eds.). (2004). Cognitive
behavioural processes across psychological disorders: A transdiagnostic approach
to research and treatment. Oxford: Oxford University Press.
Hayes, A. F. (2009). Beyond baron and kenny: statistical mediation analysis in the
new millennium. Communication Monographs, 76(4), 408e420. http://
dx.doi.org/10.1080/03637750903310360.
Hayes, A. F. (2012). Process: A versatile computational tool for observed variable
mediation, moderation, and conditional process modeling (white paper).
Retrieved from http://www.afhayes.com/public/process2012.pdf. website.
Hayes, A. (2013). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. New York: The Guilford Press.
Hedman, E., Furmark, T., Carlbring, P., Ljotsson, B., Ruck, C., Lindefors, N., et al.
(2011). A 5-year follow-up of internet-based cognitive behavior therapy for
social anxiety disorder. Journal of Medical Internet Research, 13(2), e39. http://
dx.doi.org/10.2196/jmir.1776.
Hettema, J. M. (2008). The nosologic relationship between generalized anxiety
disorder and major depression. Depression and Anxiety, 25(4), 300e316. http://
dx.doi.org/10.1002/da.20491.
Hunt, C., Slade, T., & Andrews, G. (2004). Generalized anxiety disorder and major
depressive disorder comorbidity in the national survey of mental health and
well-being. Depression and Anxiety, 20(1), 23e31.
Jacobson, N. S., & Truax, P. (1991). Clinical signiﬁcance: a statistical approach to
deﬁning meaningful change in psychotherapy research. Journal of Consulting
and Clinical Psychology, 59(1), 12e19. http://dx.doi.org/10.1037/0022006X.59.1.12.
Jones, N., Siegle, G., & Thase, M. (2008). Effects of rumination and initial severity on
remission to cognitive therapy for depression. Cognitive Therapy and Research,
32(4), 591e604. http://dx.doi.org/10.1007/s10608-008-9191-0.
Kessler, R., Andrews, G., Colpe, L., Hiripi, E., Mroczek, D., Normand, S., et al. (2002).
Short screening scales to monitor population prevalences and trends in nonspeciﬁc psychological distress. Psychological Medicine: A Journal of Research in
Psychiatry and the Allied Sciences, 32(6), 959e976.
Kroenke, K., Spitzer, R., & Williams, J. (2001). The phq-9: validity of a brief
depression severity measure. Journal of General Internal Medicine, 16(9),
606e613. http://dx.doi.org/10.1046/j.1525-1497.2001.016009606.x.
Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effects of self-focused rumination on
negative thinking and interpersonal problem solving. Journal of Personality
and Social Psychology, 69(1), 176e190. http://dx.doi.org/10.1037/0022-3514.69.
1.176.
Mahoney, A. E. J., McEvoy, P. M., & Moulds, M. L. (2012). Psychometric properties of
the repetitive thinking questionnaire in a clinical sample. Journal of Anxiety
Disorders, 26(2), 359e367.
Manicavasagar, V., Perich, T., & Parker, G. (2012). Cognitive predictors of change in
cognitive behaviour therapy and mindfulness-based cognitive therapy for
depression. Behavioural and Cognitive Psychotherapy, 40(02), 227e232. http://
dx.doi.org/10.1017/S1352465811000634.
McEvoy, P. M., Mahoney, A. E. J., & Moulds, M. L. (2010). Are worry, rumination, and
post-event processing one and the same?: development of the repetitive
thinking questionnaire. Journal of Anxiety Disorders, 24(5), 509e519. http://
dx.doi.org/10.1016/j.janxdis.2010.03.008.
McLaughlin, K. A., & Nolen-Hoeksema, S. (2011). Rumination as a transdiagnostic
factor in depression and anxiety. Behaviour Research and Therapy, 49(3),
186e193.
Newby, J. M., Mackenzie, A., Williams, A. D., Mcintyre, K., Watts, S., Wong, N., et al.
(2013). Internet cognitive behavioural therapy for mixed anxiety and depression: a randomized controlled trial and evidence of effectiveness in primary
care.
Psychological
Medicine,
1e14.
http://dx.doi.org/10.1017/
s0033291713000111.
Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the
duration of depressive episodes. Journal of Abnormal Psychology, 100(4),
569e582.
60
J.M. Newby et al. / Behaviour Research and Therapy 59 (2014) 52e60
Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and
mixed anxiety/depressive symptoms. Journal of Abnormal Psychology, 109(3),
504e511. http://dx.doi.org/10.1037/0021-843x.109.3.504.
Nolen-Hoeksema, S., Wisco, B. E., & Lyubomirsky, S. (2008). Rethinking rumination.
Perspectives on Psychological Science, 3(5), 400e424. http://dx.doi.org/10.1111/
j.1745-6924.2008.00088.x.
Norton, P. J. (2012). A randomized clinical trial of transdiagnostic cognitvebehavioral treatments for anxiety disorder by comparison to relaxation
training. Behavior Therapy, 43(3), 506e517. http://dx.doi.org/10.1016/j.beth.
2010.08.011.
Papageorgiou, C., & Wells, A. (2001). Metacognitive beliefs about rumination in
recurrent major depression. Cognitive and Behavioral Practice, 8(2), 160e164.
http://dx.doi.org/10.1016/s1077-7229(01)80021-3.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for
assessing and comparing indirect effects in multiple mediator models. Behavior
Research Methods, 40(3), 879e891.
Robinson, E., Titov, N., Andrews, G., Mcintyre, K., Schwencke, G., & Solley, K. (2010).
Internet treatment for generalized anxiety disorder: a randomized controlled
trial comparing clinician vs. Technician assistance. PLoS ONE, 5(6).
Salim, A., Mackinnon, A., Christensen, H., & Grifﬁths, K. (2008). Comparison of data
analysis strategies for intent-to-treat analysis in pre-testepost-test designs
with substantial dropout rates. Psychiatry Research, 160(3), 335e345. http://
dx.doi.org/10.1016/j.psychres.2007.08.005.
Schmaling, K. B., Dimidjian, S., Katon, W., & Sullivan, M. (2002). Response styles
among patients with minor depression and dysthymia in primary care. Journal
of Abnormal Psychology, 111(2), 350e356. http://dx.doi.org/10.1037/0021843x.111.2.350.
Schulz, K. F., Altman, D. G., Moher, D., & For The, C. G. (2010). Consort 2010 statement: updated guidelines for reporting parallel group randomised trials. PLoS
Medicine, 7(3), e1000251. http://dx.doi.org/10.1371/journal.pmed.1000251.
Segerstrom, S. C., Stanton, A. L., Alden, L. E., & Shortridge, B. E. (2003).
A multidimensional structure for repetitive thought: what's on your mind, and
how, and how much? Journal of Personality and Social Psychology, 85(5),
909e921. http://dx.doi.org/10.1037/0022-3514.85.5.909.
Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller, E., et al.
(1998). The mini-international neuropsychiatric interview (m.I.N.I): the development and validation of a structured diagnostic psychiatric interview for dsmiv and icd-10. Journal of Clinical Psychiatry, 59(Suppl. 20), 22e33.
Spek, V., Cuijpers, P., Nyklicek, I., Riper, H., Keyzer, J., & Pop, V. (2007). Internetbased cognitive behaviour therapy for symptoms of depression and anxiety: a
meta-analysis. Psychological Medicine, 37(03), 319e328. http://dx.doi.org/
10.1017/S0033291706008944.
Spitzer, R. L., Kroenke, K., Williams, J. B., & Lowe, B. (2006). A brief measure for
assessing generalized anxiety disorder: the gad-7. Archives of Internal Medicine,
166(10), 1092e1097.
Titov, N., Dear, B. F., Schwencke, G., Andrews, G., Johnston, L., Craske, M. G., et al.
(2011). Transdiagnostic internet treatment for anxiety and depression: a
randomised controlled trial. Behaviour Research and Therapy, 49(8), 441e452.
http://dx.doi.org/10.1016/j.brat.2011.03.007.
Watkins, E. R. (2008). Constructive and unconstructive repetitive thought. Psychological Bulletin, 134(2), 163e206. http://dx.doi.org/10.1037/0033-2909.134.2.163.
Watkins, E. R. (2009). Depressive rumination and co-morbidity: evidence for
brooding as a transdiagnostic process. Journal of Rational-Emotive & CognitiveBehavior Therapy, 27(3), 160e175.
Watkins, E., & Moulds, M. L. (2005). Positive beliefs about rumination in depression
e a replication and extension. Personality and Individual Differences, 39(1),
73e82. http://dx.doi.org/10.1016/j.paid.2004.12.006.
Watkins, E., Scott, J., Wingrove, J., Rimes, K., Bathurst, N., Steiner, H., et al. (2007).
Rumination-focused cognitive behaviour therapy for residual depression: a case
series. Behaviour Research and Therapy, 45(9), 2144e2154. http://dx.doi.org/
10.1016/j.brat.2006.09.018.
Watkins, E. R., Taylor, R. S., Byng, R., Baeyens, C., Read, R., Pearson, K., et al. (2012).
Guided self-help concreteness training as an intervention for major depression
in primary care: a phase ii randomized controlled trial. Psychological Medicine,
42(07), 1359e1371. http://dx.doi.org/10.1017/S0033291711002480.
Watson, D. (2005). Rethinking the mood and anxiety disorders: a quantitative hierarchical model for dsm-v. Journal of Abnormal Psychology, 114(4), 522e536.
Wells, A. (1995). Meta-cognition and worry: a cognitive model of generalized
anxiety disorder. Behavioural and Cognitive Psychotherapy, 23(3), 301e320.
Wells, A., & Matthews, G. (1994). Attention and emotion: A clinical perspective.
Hillsdale, NJ, England: Lawrence Erlbaum Associates, Inc.
Wells, A., & Matthews, G. (1996). Modelling cognition in emotional disorder: the sref model. Behaviour Research and Therapy, 34(11e12), 881e888. http://
dx.doi.org/10.1016/S0005-7967(96)00050-2.
World Health Organization. (2008). The global burden of disease: 2004 update.
Geneva, Switzerland: WHO Press.

Download Report