Internet Interventions 2 (2015) 272–282

Contents lists available at ScienceDirect

Internet Interventions

journal homepage: www.invent-journal.com/

The efficacy of internet-delivered treatment for generalized anxiety

disorder: A systematic review and meta-analysis
Derek Richards a,b,⁎, Thomas Richardson c,d, Ladislav Timulak b, James McElvaney b
a
SilverCloud Health, The Priory, John Street West, Dublin, Ireland
b
School of Psychology, University of Dublin, Trinity College, Dublin, Ireland
c
Professional Training Unit, School of Psychology, University of Southampton, Southampton, UK
d
Mental Health Recovery Team South, Solent NHS Trust, Portsmouth, UK

a r t i c l e i n f o a b s t r a c t

Article history: Generalized Anxiety Disorder (GAD) is typically considered a chronic condition characterized by excessive worry.
Received 10 May 2015 Lifetime prevalence is 4.3–5.9%, yet only a small percentage seeks treatment. GAD is treatable and in recent years
Received in revised form 9 July 2015 internet-delivered treatment interventions have shown promise. This paper aims to systematically search for lit-
Accepted 12 July 2015
erature on internet-delivered psychological interventions for the treatment of GAD and conduct a meta-analysis
Available online 19 July 2015
to examine their efficacy. The purpose of the paper is to inform the community of researchers, program devel-
Keywords:
opers and practitioners in internet delivered interventions of the current state-of-the-art and research gaps
Generalized anxiety disorder that require attention. A systematic search of the literature was conducted to find all studies of internet-delivered
Anxiety treatments for GAD (N = 20). Using Review Manager 5 all Randomized Controlled Trials (RCTs; n = 11) that met
Internet interventions our established eligibility criteria were included into a meta-analysis that calculated effect sizes via the standard-
Efficacy ized mean difference. Compared to the waiting-list controls, the results demonstrate positive outcomes for GAD
Meta-analysis symptoms (d = −0.91) and its central construct of pathological worry (d = −0.74). The meta-analysis supports
Systematic review the efficacy of internet-delivered treatments for GAD including the use of disorder-specific (4 studies) and
transdiagnostic treatment protocols (7 studies). Caution is advised regarding the results as the data is limited
and highly heterogeneous, but revealing of what future research might be needed.
© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction (Kessler et al., 2005a). Research suggests that GAD is a chronic and en-
during condition (Angst & Vollrath, 1991; Grant et al., 2005). Further-
Generalized Anxiety Disorder (GAD) is characterized by excessive more, comorbidity is as high as 90%, with 70% being diagnosed with
anxiety and worry, which the sufferer describes as difficult to control, oc- comorbid depression, over 55% with any other anxiety disorder and
curring more days than not for a period of at least six months (American 48% with a somatoform disorder (Carter et al., 2001). Around 50% of pa-
Psychiatric Association [APA], 2013). Other symptoms of GAD include tients with GAD have also a personality disorder, most commonly
restlessness, being easily fatigued, difficulty concentrating, irritability, avoidant and dependent personality disorder (Sanderson et al., 1994).
muscle tension, and sleep disturbance. GAD is one of the most prevalent Depression is commonly shown to follow GAD (Kessler et al., 2004),
anxiety disorders (Kessler et al., 2005a; Kessler et al., 2005b; Narrow suggesting that chronic GAD may start the onset of depression in
et al., 2002). Its one-year prevalence in community samples in the US is some cases (Barlow, 2002).
around 3% and its lifetime prevalence around 5% (Blazer et al., 1991; People with GAD experience significant impairment in quality of life
Kessler et al., 2005a; Kessler et al., 2005b; Wittchen, 2002). Studies from (Loebach Wetherell et al., 2004; Massion et al., 1993). GAD negatively
other countries revealed roughly similar figures (Bijl et al., 1998; impacts the individual's general sense of well-being and life satisfaction
Faravelli et al., 1989; Hunt et al., 2002; Jenkins et al., 1997). GAD patients and specifically occupational and family satisfaction (Stein & Heimberg,
typically present in primary care settings, where the reported prevalence 2004). GAD represents a significant cost to society due to disability, de-
is up to 8% (Kroenke et al., 2007; Roy-Byrne & Wagner, 2004). creased work productivity and increased use of health care services
Evidence from retrospective accounts suggest that people with GAD (Wittchen, 2002).
will have their first episode by age 31, with a quarter having their first
episode by age 20, with an early onset in childhood or adolescence 2. GAD and its treatment

As is the case with other anxiety disorders, cognitive-behavioral

⁎ Corresponding author at: School of Psychology, Trinity College, Dublin, Ireland. therapy (CBT), a form of psychological therapy, is the treatment that is

http://dx.doi.org/10.1016/j.invent.2015.07.003
2214-7829/© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 D. Richards et al. / Internet Interventions 2 (2015) 272–282 273

routinely considered for GAD (National Institute for Health and Clinical technology-assisted CBT for anxiety concluded that the field was in its
Excellence, 2011). CBT for GAD is well studied and is shown to be more infancy but that existing research was promising and suggested that
effective than wait-list, non-specific control conditions or treatment as technology-based delivery may be efficacious and cost-effective. Reger
usual (Borkovec & Ruscio, 2001; Hunot et al., 2007; National Institute and Gahm's (2009) meta-analysis concluded that the data supported
for Health and Clinical Excellence, 2011). Several cognitive-behavioral the use of such delivery systems and that the results are superior to
models of generalized anxiety disorder exist (Brown et al., 2001; waiting-list or placebo. The study did not review any internet interven-
Dugas & Robichaud, 2007). In general, these models assume that people tions for the treatment of GAD (Reger & Gahm, 2009); there simply
with GAD had early experiences of uncontrollability (Brown et al., 2001) were none published at the time. A similar meta-analysis by Cuijpers
or have intolerance of uncertainty on the basis of negative belief (Dugas et al. (2009) found a large effect size (d = 1.08) for the active conditions
& Robichaud, 2007). The worry in GAD aims at avoiding future aversive compared to the controls. The authors concluded in favor of the poten-
events (Borkovec, 1994; Brown et al., 2001) which brings temporary re- tial of computer-aided delivery of treatments for anxiety disorders
lief, but also inhibition of emotional processing and maintenance of (Cuijpers et al., 2009). The meta did not include studies for generalized
anxiety-producing thinking and behaviors (Brown et al., 2001). More anxiety disorder. Andrews et al. (2010) published a meta-analysis dem-
cognitive and meta-cognitive models of GAD also stress positive beliefs onstrating that computerized CBT was superior to outcomes from con-
about worry's protective function (Wells, 1999). trol groups. In an analysis of 22 studies of comparisons with a control
CBT for GAD includes a number of specific components such as cogni- group they reported a post-treatment effect size of d = 1.12 for the
tive restructuring, behavioral exposure to feared consequences, worry ex- studies that examined GAD (Andrews et al., 2010).
posure (staying with feared outcomes), relaxation training, worry To date, the data available for the relevance of internet-delivered
behavior prevention and problem solving (Borkovec & Ruscio, 2001; treatments on outcomes specifically in GAD-diagnosed subjects is
Brown et al., 2001; Covin et al., 2008; Dugas & Robichaud, 2007). Their scarce. In recent years, principally using internet-delivery, other studies
main rationale is that the patient overcomes emotional avoidance and have been published. A recent meta-analysis by Cuijpers et al. (2014)
learns that anxiety is not debilitating, but manageable and recedes after examined the effectiveness of psychological therapy for GAD. While it
time. Recently, transdiagnostic CBT protocols for depressive and anxiety included studies using internet interventions it was not their primary
disorders have been proposed that also focus on features relevant to focus.
GAD such as emotional avoidance (Barlow et al., 2011). Proponents of In 2013, Cochrane published a review of media-delivered cognitive
transdiagnostic interventions argue that the similarities between the anx- behavior therapy and behavior therapy (self-help) for anxiety disorders
iety disorders outweigh their individual differences and they can respond in adults (Mayo-Wilson & Montgomery, 2013). Some of the studies in
to common therapeutic procedures (Allen et al., 2007). that review we include here also. The other studies they included for
GAD are unpublished data from Kiely (2002), Houghton (2008), and
3. Access to treatment: the evolution of high and low-intensity Shoenberger (2008). They included Bowman (1997), but the media
interventions used was worksheets on paper (not computer or internet-delivered)
and lastly Rosmarin (2010), which included a sub-clinical anxiety
Healthcare providers are increasingly faced with a discrepancy be- group, not GAD symptom-specific group. Their search period ended Jan-
tween the burden of mental health conditions and the availability of uary 1 2013 and further (n = 4) studies have been published since that
cost-effective psychological treatments (Kohn et al., 2004). It has been time. Similarly, Christensen et al. (2014) included a search period of 18
estimated that upwards of 70% of people with anxiety disorders go un- months from 2012 to June 2013 and included two studies for GAD, but
treated every year (Andrews et al., 2001; Lepine, 2002). There are mul- since that time other studies have been published.
tiple barriers to accessing treatment, including waiting-lists, costs, The current study therefore aimed to be more specific and systemat-
distance from service locations, negative perception of treatments, and ically review and conduct a meta-analysis of internet-delivered psycho-
personal stigma (Kohn et al., 2004; Mohr et al., 2010). logical therapy for GAD compared to waiting-list control groups. The
In recent years a model of stepped-care has evolved, involving purpose of the paper is to inform the community of researchers, pro-
high-intensity (e.g., one-to-one therapy) and low-intensity gram developers and practitioners in internet-delivered interventions
(e.g., bibliotherapy, internet-delivered treatments) interventions of the current state of the art and research gaps that require attention.
(Bower & Gilbody, 2005). Low-intensity internet-delivered inter- The paper presents a comprehensive search of the literature, an effort
ventions have the potential to extend access and reduce costs and to gather discrete data on subjects, and a detailed focus on the efficacy
possibly can overcome some of the barriers mentioned above. of internet interventions on GAD specific and some co-morbid (depres-
Several studies have reported positive outcomes for internet- sion, distress, disability and quality of life) symptoms.
delivered treatments for social phobia, spider phobia, flight and other
phobias, panic disorder, obsessive-compulsive disorder (OCD), post-
traumatic stress disorder (PTSD), stress-related anxiety, trauma, de- 5. Method
pression and generalized anxiety disorder (GAD) (Cuijpers et al.,
2009; Reger & Gahm, 2009; Richards & Richardson, 2012). 5.1. Literature search and selection of studies
Internet-delivered cognitive behavior therapy treatment protocols
have included disorder-specific treatments and transdiagnostic treat- The aim of our literature search was to find all studies that related to
ments that aim to treat the common elements and symptoms for anxi- internet-delivered treatment protocols for GAD, including disorder-
ety disorders in general (Andersson et al., 2012; Bell et al., 2012; specific protocols and more recent transdiagnostic protocols. During
Carlbring et al., 2011; Johnston, Titov, Andrews, Spence, & Dear, 2011; June 2013, we selected three prominent databases (Embase, PubMed,
Newby et al., 2013; Paxling et al., 2011; Robinson et al., 2010; Titov and PsychINFO including PsychARTICLES) as our search arena. After ini-
et al., 2010; Titov et al., 2009; Titov et al., 2011). Few internet-delivered tial experimentation with several search phrases (online delivered
treatments have integrated other therapeutic practices such as brief treatments for anxiety/generalis[z]ed anxiety, web-based treatment/in-
psychodynamic therapy (Andersson et al., 2012). terventions for anxiety, among others) that were derived from the au-
thors' experiences in internet-delivered treatments and also from
4. Other reviews and meta-analyses known studies, we decided on the use of three key search phrases that
we were confident would yield the relevant literature. They were ‘inter-
A number of reviews and meta-analyses of this area have been pub- net treatment for generalized anxiety disorder’ and ‘internet treatment
lished. An early narrative review (Przeworski & Newman, 2006) of for generalized anxiety disorder’ and ‘internet treatment for anxiety’.
274 D. Richards et al. / Internet Interventions 2 (2015) 272–282

We used the three search phrases across the three databases chosen, So as to include discrete outcomes for patients undergoing treatment
culminating in a total of nine searches. for GAD, studies that employed transdiagnostic anxiety treatment pro-
Initially, search results were excluded at title depending on their rel- tocols had to discriminate outcomes for the different anxiety disorders
evance, thereafter abstracts were read and further papers excluded. to be included. In some cases where a transdiagnostic protocol was
Lastly the remaining papers were read fully and excluded if they were employed and/or outcomes from several anxiety disorders reported in
not eligible (for eligibility criteria see below). Finally, the reference total for participants, we contacted authors to get the discrete outcomes
lists of accepted papers and other reviews and meta-analysis data for the GAD diagnosed subjects. All the studies were published in
(Andrews et al., 2010; Cuijpers et al., 2009; Przeworski & Newman, peer-reviewed journals in English and included reliable and valid mea-
2006; Reger & Gahm, 2009) were checked for further relevant papers. sures for the assessment of outcomes, such as Generalized Anxiety
The process was conducted by the first two authors (DR, TR) and any Disorder-Q-IV (Newman et al., 2002) and Generalized Anxiety
disagreements that arose were discussed until a final decision was Disorder-7 (Spitzer et al., 2006) and the Penn State Worry Question-
reached. naire (Meyer et al., 1990).
Eligibility criteria was established to include studies that were ran- Post data-analysis and manuscript preparation, we carried out a fur-
domized controlled trials of an internet-delivered intervention com- ther search and due to time lapsed, included the original search arena
pared to a waiting-list control. The studies were based on adult (18 + and added the Cochrane database, and then also did a search by cite.
years) samples that had a clinical diagnosis of GAD, whom may have The search yielded one further study, but it did not meet our eligibility
had comorbidity with depression and/or impairment in functioning. criteria for inclusion in the meta-analysis (Boettcher et al., 2014).

Fig. 1. Results from the systematic search.

Table 1
Studies included in the systematic review.

Study Participants Sample Design Intervention Support Measures Country

*Andersson et al. Telephone administered SCID-I Community (n = 81) RCT: ICBT Therapistv (final year clinical psychology PSWQ Sweden
(2012) diagnosis for GAD ICBT: 27 IPDT — SUBGAP trainees and one licensed psychologist) GAD-IV
IPDT: 27 8 content modules/8 weeks MADRS-S
WL: 27 BAI
BDI-II
QOLI
STAI
*Bell et al. (2012) Referrals to anxiety disorder unit. Clinical (n = 83) RCT: CCBT: 4 sessions of CBT within Research assistant (not clinical) telephone WASA New Zealand
SCID-I administered in person. CCBT: 7 12 weeks call every 2 weeks for compliance. GADI
WL: 7 WL PSWQ
BDI-II
*Berger et al. (2013) Telephone administered SCID-I Community (n = 132: sample RCT ICBT — 8-content modules/8 Masters level therapists (final year clinical BAI Switzerland
diagnosis for GAD recruited from Germany, TAiCBT: 44 weeks psychologists, a qualified psychologist, and a BDI-II
Switzerland and Austria) STiCBT: 44 qualified CBT therapist) weekly written GSI (BSI)
WL: 44 feedback. SPS
SIAS
ACQ
BSQ

D. Richards et al. / Internet Interventions 2 (2015) 272–282

MIA
MIB
PSWQ
Boettcher et al. Telephone administered SCID-I Community (n = 91) RCT Internet-based mindfulness None for intervention group. Supervised BAI Sweden
(2014) diagnosis for anxiety disorder IBMT: 45 treatment for anxiety discussion forum for WL group. BDI-II (recruitment through website
WL: 46 disorders (transdiagnostic) 8 ISI and newspaper adverts)
modules/8 weeks QOLI
Carlbring et al. In person administered SCID for Community RCT Individually tailored CBT for Advanced MSc Clinical Psychology students. BAI Sweden
(2011) diagnosis of an anxiety disorder (n = 54) iCBT: 27 comorbid anxiety disorders Weekly e-mail feedback. CORE-OM (recruitment through website, ra-
Control: 27 (and depression). MADRS-S dio interviews, and newspaper
6–10 modules (out of 16)/10 QOLI adverts)
weeks.
Craske et al. (2011) MINI for diagnosis of 1 or more anxiety Clinical (n = 1004) RCT iCBT disorder-specific (PD, Practitioners (social workers, nurses, MSc PDSS-SR USA
disorders. iCBT: 503 PTSD, GAD, SAD) and PhD-level psychologists) worked GADSS (recruitment through referral in
(270 GAD) 8 modules/10–12 weeks collaboratively with Ps in person as they SPIN primary care setting)
Control: 501 completed the program. PCL-C
(279 GAD)
Dear et al. (2011) MINI telephone-administered Community (n = 32) Open trial iCBT transdiagnostic Clinical Psychologist DASS-21 Australia
(single-sample) depression and anxiety Weekly telephone or text-based support PHQ-9
disorders PSWQ
5 modules/8 weeks SIAS6/SPS6
PDSS-SR
GAD-7
K-10
SDS
NEO-FFI-N
Draper et al. (2008) In person administered SCID for Clinical (n = 3) Multiple case series iCBT (GAD-specific) Encouragement provided by “occasional” PSWQ Australia
diagnosis of GAD 11 modules/11 weeks telephone contact (do not specify who GADQ-IV
provided contact) MCQ-30
*Johansson et al. Telephone administered MINI Community (n = 100) RCT: Based on APT model Masters level therapists. Weekly written PHQ-9 Sweden
(2013) interview diagnosis for depression and IPDT: 50 8 content modules/10 weeks feedback. GAD-7
anxiety disorder WL: 50 EPS-25
FFMQ
*Johnston et al. Telephone administered MINI diagnosis Community (n = 139) RCT: Anxiety Program of 8 content Weekly telephone or email contact from GAD-7 Australia
(2011) for GAD, social phobia, or panic ICBT-CL: 46 modules/10 weeks either clinician or coach DASS-21
disorder ICBT-CO: 47 PSWQ
WL: 46 SIAS-6
SPS-6
PDSS-SR
PHQ-9
SDS

275
(continued on next page)
 276
Table 1 (continued)

Study Participants Sample Design Intervention Support Measures Country

Klein et al. (2011) “e-PASS” online Ax (540 items Community (n = 225) Quasi-experimental 5 iCBT programs specific to e-mail support. K-6 Australia
corresponding to DSM-IV-TR criteria) (naturalistic GAD, PD/A, OCD, PTSD, or Therapists or postgraduate psychology e-PASS
to refer clients to program appropriate participant choice). SAD. students.
to their difficulties. GAD = 88 12 modules/12 weeks
Mewton et al. (2012) Referred by practitioners. Clinical (n = 588) Naturalistic iCBT GAD specific 6 Nature of support not specified. Prescribing GAD-7 Australia
single-sample modules/time not specified practitioner received updates on client K-10
progress WHODAS
*Newby et al. (2013) telephone administered MINI to Community (n = 109) RCT: Worry and sadness program – Regular therapist contact for first two PHQ-9 Australia
confirm diagnosis of GAD and/or MDD ICBT: 49 6 content modules/10 weeks modules, as needed subsequently. GAD-7
WL: 60 (telephone/e-mail) K-10
WHODAS-II
BDI-II
PSWQ
NEO-FFI-N
*Paxling et al. (2011) Telephone administered SCID-I Community (n = 89) RCT: ICBT – 8 content modules/8 Therapist — weekly email feedback PSWQ Sweden
diagnosis for GAD ICBT: 44 weeks GAD-IV
WL: 45 STAI
BAI
BDI

D. Richards et al. / Internet Interventions 2 (2015) 272–282

MDRS
QOLI
*Robinson et al. Telephone administered MINI diagnosis Community (n = 150) RCT: ICBT 6 content modules/10 Weekly supportive e-mail or telephone PSWQ Australia
(2010) for GAD ICBT-TA:50 weeks contact from Therapist (clinical psychologist) GAD-7
ICBT-CA: 51 or technician (clinic administrator). K-10
WL: 49 SDS
PHQ-9
Sunderland et al. Primary diagnosis of GAD or Clinical (n = 663) Naturalistic iCBT for GAD Progress overseen by prescribing clinician (level K-10 Australia
(2012) depression, referred by GP/mental single-sample (and iCBT for depression) and method of contact not specified) PHQ-9
health professional. iCBT dep: 302 6 modules/10 weeks GAD-7
iCBTanx: 361
*Titov et al. (2009) Telephone administered MINI diagnosis Community (n = 48) RCT: ICBT 6 content modules/9 Weekly therapist support (clinical GAD-7 Australia
for GAD ICBT: 25 weeks psychologist) PSWQ
WL: 23 PHQ-9
K-10
SDS
*Titov et al. (2010) Telephone administered MINI diagnosis Community (n = 86) RCT: ICBT Anxiety 6 content Weekly text-based and/or telephone contact GAD-7 Australia
for GAD, social phobia, panic disorder ICBT: 42 modules/8 weeks from therapist (clinical psychologist) PSWQ
WL: 44 SPSQ
PDSS-SR
PHQ-9
K-10
SDS
DASS-21
*Titov et al. (2011) Telephone administered MINI diagnosis Community (n = 78) RCT: Wellbeing program — 8 Weekly text-based and/l or telephone from DASS-21 Australia
for anxiety disorder or depression ICBT: 39 content modules/10 weeks therapist (clinical psychologist) PHQ-9
WL: 38 PSWQ
SP-12
PDSS-SR
GAD-7
K-10
SDS
Zou et al. (2012) Telephone administered MINI for Older adult community Single sample iCBT for anxiety disorders Clinical psychologist. Weekly telephone or GAD-7 Australia
diagnosis of an anxiety disorder (n = 22) (transdiagnostic) e-mail support. DASS-21
5 modules/8 weeks PHQ-9
SDS
K-10

Note. * = indicates studies included in the meta-analysis; SCID-I = Structured Clinical Interview for DSM-IV Axis I Disorders; DSM-IV = Diagnostic and Statistical Manual for Mental Health Disorders-IV; MINI = International Neuropsychiatric Interview;
GAD = Generalized Anxiety Disorder; MDD = Major Depressive Disorder; RCT = Randomized Controlled Trial; ICBT = Internet Cognitive Behavior Therapy; IPDT = Internet Psychodynamic Therapy; WL = Waiting-List; APT = Affect-Phobia Therapy;
TAiCBT = Tailored internet cognitive behavior therapy; STiCBT = Standardized internet cognitive behavior therapy; PSWQ = Penn State Worry Questionnaire; GAD-IV = Generalized Anxiety Disorder-IV; MADRS-S = Montgomery–Åsberg Depression
Rating Scale; BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory-II; QOLI = Quality of Life Inventory; STAI = State-Trait Anxiety Inventory; PDSS-SR = Panic Disorder Severity Scale; K-10 = Kessler-10; SDS = Sheehan Disability Scale; PHQ-
9 = Patient Health Questionnaire; DASS-21 = Depression Anxiety Stress Scale-21; SPSQ = Social Phobia Screening Questionnaire; WHODAS-II = World Health Organization Disability Assessment Schedule II; SIAS-6 = Social Interaction Anxiety Scale; SPS-
6 = Social Phobia Scale — Short Form; GADI = Generalized Anxiety Disorder Assessment Inventory; WSAS = Work & Social Adjustment Scale; LSAS = Liebowitz Social Anxiety Scale; FQ = Fear Questionnaire; ACQ = Agoraphobic Cognitions Questionnaire;
BSQ = Body Sensations Questionnaire; MIA = Mobility Inventory for Agoraphobia; TA = Technician Assisted; CA = Clinician Assisted; CL = Clinician-supported; CL = Coach-supported.
 D. Richards et al. / Internet Interventions 2 (2015) 272–282 277

5.2. Meta-analysis method symptoms. In conclusion, we received discrete outcomes for GAD diag-
nosed subjects from the authors of 7 transdiagnostic studies (Bell et al.,
All studies included were assessed for data which could be included 2012; Berger et al., 2013; Johansson et al., 2013; Johnston et al., 2011;
in a meta-analysis of effects sizes at post-treatment comparing internet- Newby et al., 2013; Titov et al., 2010; Titov et al., 2011).
delivered therapy to waiting-list controls. Any variable (e.g. symptoms We had to exclude the use of data from 1 study (the authors could
of generalized anxiety and worry and co-morbid depression) which not supply the discrete data for GAD diagnosed participants), which
was reported by two or more studies was analyzed. We included the meant that we included data from 11 studies in this meta-analysis.
analysis of depression symptoms, as typically these are measured in Fig. 1 shows the results of the systematic search and the reasons for
both single diagnosis focused studies and transdiagnostic. Eleven stud- exclusion.
ies reported data for both active interventions and control groups (all
were waiting-list controls) which could be used. If different measures 6.2. Overview of the studies included
were used these were combined if they measured the same construct.
For example studies reporting scores on the PHQ-9 (Spitzer et al., Selected characteristics of the studies can be found in Table 1. In total
1999) and Beck Depression Inventory-II (Beck et al., 1996) were we were able to include 11 distinct studies that reported on outcomes
combined into a ‘self-reported depression’ category. Three studies from disorder-specific or transdiagnostic treatments for generalized
(Andersson et al., 2012; Johnston et al., 2011; Robinson et al., 2010) anxiety disorder. In the 11 studies 771 participants were included either
had two active intervention conditions as they compared different as part of active treatments (n = 371) or as waiting-list controls (n =
types of internet-delivered therapy. Data from both intervention condi- 400).
tions was included in the meta-analysis. We decided to exclude comor- The majority of participants were recruited via a website where they
bid anxiety disorders in the analysis due to the fact that many of the visited or had already registered their interest prior to the trial. It seems
studies were transdiagnostic, multiple anxiety disorders focused, and that all of the Australian studies recruited through a website (www.
interventions. virtualclinic.org.au) alongside a community-based newspaper advert
All available data was continuous and was therefore analyzed using in one case (Johnston et al., 2011). Most other studies also relied on
standardized mean difference (Cohens d), weighted by sample size via a self-recruitment through websites advertising the studies and adverts
random effects model with 95% confidence interval to compare post- in local community newspapers. In one case recruitment was from sev-
treatment scores between waiting-list controls and active samples. eral countries (Berger et al., 2013). Samples were community-based,
There was insufficient data to analyze outcomes at follow-up. The prin- apart from Bell et al. (2012) who recruited from a clinical population,
cipal measures for GAD symptoms were the Generalized Anxiety and ranged in size from 48 participants (Titov et al., 2011) to 150 partic-
Disorder-Q-IV (Newman et al., 2002) and Generalized Anxiety ipants (Robinson et al., 2010), the mean sample size across the 11 stud-
Disorder-7 (Spitzer et al., 2006), and for pathological worry the measure ies was 99 participants.
was the Penn State Worry Questionnaire (Meyer et al., 1990). After completing initial screening and intake questionnaires, all of
We also calculated homogeneity of \effect size using the I2-statistic the studies administered either the Structured Clinical Interview for
that indicates heterogeneity as a percentage. A value of 0% indicates DSM-IV Axis I Disorders (SCID-I; First et al., 1997), the Mini Internation-
no observed heterogeneity, larger values represent increases in hetero- al Neuropsychiatric Interview Version 5.0.0 (MINI; Sheehan et al.,
geneity, for instance 25% is considered low, 50% as moderate and 75% 1998), or an interview based on the MINI (Johansson et al., 2013) to es-
considered high heterogeneity (Higgins et al., 2003). tablish a formal diagnosis of GAD. The majority administered the inter-
view over phone while one administered the interview in person
5.2.1. Assessment of the quality of the included studies (Johansson et al., 2013; Newby et al., 2013). All of the studies employed
To assess the validity of the included studies we employed the risk of what can be considered robust and usual measures to assess outcomes
bias assessment developed by Higgins and Green (2009) for the from treatment. These included for the most part the Generalized Anxiety
Cochrane Collaboration. The first two authors (DR, TR) assessed the Disorder Inventory-7 (GAD-7), and the Penn State Worry Questionnaire
studies on four key questions; 1. Was the allocation sequence adequate- (PSWQ). Berger et al. (2013) argued that because not all participants suf-
ly generated (Selection Bias)? 2. Was allocation adequately concealed fered from the same primary anxiety disorder, disorder-unspecific mea-
(Selection Bias)? 3. Was knowledge of the allocated interventions ade- sures were employed to assess primary outcomes, namely the Beck
quately prevented during the study (Performance Bias)? and 4. Were Anxiety Inventory (BAI; Beck & Robert, 1993). The study did include sec-
outcome assessments adequately managed (Detection Bias)? ondary outcome measures for specific anxiety disorders, such as the So-
cial Phobia Scale (SPS). Something similar is witnessed in some studies
6. Results that delivered a transdiagnostic treatment, a mix of primary and second-
ary measures were used to discretely assess outcomes among the anxiety
6.1. Selection and inclusion of studies disorders (Johnston et al., 2011; Titov et al., 2010; Titov et al., 2011).
Four of the studies can be considered disorder-specific, whose inter-
From the database searches twenty studies (n = 20) were included ventions directly address generalized anxiety disorder (Andersson et al.,
into the systematic review. Twelve of these studies were selected for in- 2012; Paxling et al., 2011; Robinson et al., 2010; Titov et al., 2009). The
clusion in the meta-analysis. Four studies were GAD specific samples remaining seven studies were transdiagnostic in that they were direct-
and interventions (Andersson et al., 2012; Paxling et al., 2011; Robinson ed at either multiple anxiety disorders (Berger et al., 2013; Johnston
et al., 2010; Titov et al., 2009), the remainder were transdiagnostic. One et al., 2011; Titov et al., 2010) or anxiety disorders and depression
author alerted us to another study that we missed in our search in June (Johansson et al., 2013; Newby et al., 2013; Titov et al., 2011). The treat-
2013 as the paper was published in July 2013 (Johansson et al., 2013); ment intervention delivered in 9 of the 11 studies was based on cogni-
we also came upon a newly published paper in our second search before tive and behavioral principles. Two studies employed a psychodynamic
submitting the manuscript for publication (Berger et al., 2013), we decid- intervention (Andersson et al., 2012; Johansson et al., 2013). All of the
ed to include these studies as they complied with our established eligibil- studies involved an individual treatment format and the treatments
ity criteria. were predominately delivered over 8 sessions of content on a weekly
One study conducted telephone MINI interviews to confirm diagnosis basis (Andersson et al., 2012; Berger et al., 2013; Johnston et al., 2011;
of GAD and/or MDD (Major Depressive Disorder) (Newby et al., 2013). Paxling et al., 2011; Titov et al., 2011), or with an extended delivery
We decided to include the data from this study as all had primary diagno- time of 10 weeks (Johnston et al., 2011, Titov et al., 2011, Johansson
sis of GAD or had a diagnosis of MDD with significant subthreshold GAD et al., 2013). Four interventions were delivered in 6 modules of content
278 D. Richards et al. / Internet Interventions 2 (2015) 272–282

over 8 to 10 weeks (Newby et al., 2013; Robinson et al., 2010; Titov 6.5. Meta-analysis results: effects of internet-delivered interventions
et al., 2010; Titov et al., 2009). The GAD treatment in Bell et al. (2012)
consisted of 4 lessons to be completed within 12 weeks. Results from the meta-analysis are shown in Table 2. Sample sizes
In line with best practice in internet-delivered treatments for anxi- for the individual analyses ranged from 66 to 344 in the treatment
ety and depression (Newman et al., 2011; Richards & Richardson, conditions. There were statistically significant improvements for
2012) support for participants was provided in most of the treatment internet-delivered interventions compared to waiting-list controls on
conditions. Support was provided by therapists in all conditions at var- self-reported GAD symptoms (d = − 0.91; CI: 1.25–0.56; n = 8) and
ious stages of training in clinical psychology, masters or doctorate pathological worry (d = −0.74; CI: 0.96–0.52; n = 10), both yielding
courses. In some cases, qualified and experienced therapists/clinical what can be considered large effects (Cohen, 1988). Similar statistically
psychologists provided participant support. In the case of Johnston significant large effects can be noted for the active treatments compared
et al. (2011) the coach supporter was a graduate psychologist with no to waiting-list controls for comorbid anxiety (d = −0.57), depression
further postgraduate training. Bell et al. (2012) did not provide thera- (d = −0.63), distress (d = −0.91), disability (d = −0.77), and quality
peutic support for their participants; a research assistant provided of life (d = 0.38). Figs. 3 and 4 display forest plots for the primary out-
short, highly structured phone calls every 2 weeks. come variables of GAD and Worry. Figs. 5 and 6 display funnel plots for
these variables. The funnel plot for Worry (Fig. 6) is relatively symmet-
rical suggesting no clear publication bias. However GAD (Fig. 5) is some-
6.3. Overview of the studies excluded what asymmetrical suggesting some publication bias.
There was high heterogeneity observed for GAD (I2 = 77%) and de-
Studies that did not meet our eligibility criteria as outlined in the pression (I2 = 68%), moderate heterogeneity observed for worry (I2 =
method were therefore excluded from the meta-analysis. Selected char- 46%), and distress (I2 = 39%), for the other variables the observed het-
acteristics of the studies can be found in Table 1. In these studies, 2682 erogeneity was not significant. Given the moderate to high heterogene-
participants were included either as part of active treatments (n = ity observed for some variables it would suggest significant variance in
2108) or as waiting-list controls (n = 574). Participants were recruited the distribution of the effect sizes reported. However, with the removal
from the community through websites (Boettcher et al., 2014; Carlbring of the psychodynamic studies (Andersson et al., 2012; Johansson et al.,
et al., 2011; Dear et al., 2011; Klein et al., 2011), or from clinical popula- 2013) the results indicate a reduction in heterogeneity and an increase
tions via referral from GPs or mental health practitioners (Craske et al., in effect of the two main constructs, namely generalized anxiety symp-
2011; Draper et al., 2008; Mewton et al., 2012; Sunderland et al., 2012; toms (I2 = 49%; d = 1.19) and worry (I2 = 17%; d = 0.87). There was
Zou et al., 2012). Sample size ranged from 3 (Draper et al., 2008) to statistically significant variation for GAD, worry, and depression, how-
1,004 (Craske et al., 2011). Most studies used one or more self-report ever, changing from a random to fixed-effects model had little impact
measures to assess anxiety. Most studies included a measure that on effect sizes, suggesting that heterogeneity for these variables was
specifically assessed GAD, such as the Generalized Anxiety Disorder not problematic.
Inventory-7 (GAD-7) (Dear et al., 2011; Mewton et al., 2012; Three studies (Johansson et al., 2013; Newby et al., 2013; Titov et al.,
Sunderland et al., 2012; Zou et al., 2012); and the Penn State Worry 2011) included participants with depression and anxiety disorders,
Questionnaire (PSWQ) (Dear et al., 2011; Draper et al., 2008). however, with the exclusion of these subjects with depression and anx-
Five studies evaluated GAD-specific internet-delivered cognitive-be- iety, the effect size for depression remained the same d = − 0.63
havioral therapy (iCBT) interventions (Craske et al., 2011; Draper et al., (−1.03, −0.91).
2008; Klein et al., 2011; Mewton et al., 2012; Sunderland et al., 2012). Sub-group analyses were conducted to compare studies which were
Two studies looked at transdiagnostic internet-delivered CBT programs: GAD specific to studies which were transdiagnostic (specifically
one for anxiety disorders (Zou et al., 2012), and one for comorbid anxi- transdiagnostic or included comorbid depression or other anxiety disor-
ety and depression (Dear et al., 2011). Carlbring et al. (2011) evaluated a der). For GAD subjects, effect sizes were similar for GAD-specific
tailored internet-delivered CBT approach where clients are assigned 6– (d = − 0.81; CI: − 1.27, − 0.35, n = 4, p b .001) and transdiagnostic
10 out of a possible 16 modules for anxiety disorders based on their spe- (d = −.91; CI: −1.25, −0.56, n = 4, p b .001). The difference between
cific diagnosis or diagnoses. The final study evaluated a transdiagnostic these subgroups was not statistically significant: χ2 = 0.34, df = 1,
internet-based mindfulness treatment for the anxiety disorders. The in- p N .05. For worry the effect sizes were also similar for GAD-specific
tervention programs in the studies ranged from 5 modules over 8 weeks (d = − 0.68; CI: − 0.97, − 0.38, n = 5, p b .001) and transdiagnostic
to 12 modules over 12 weeks. In line with best practice, most studies in- (d = −0.77; CI: −1.12, −0.42, n = 4). The difference between these
cluded support from a practitioner, by telephone, e-mail or instant mes- subgroups was not statistically significant: χ2 = 0.16, df = 1, p N .05.
saging. Outcomes from across the studies are positive for the clinical
benefit of internet-delivered interventions for GAD. 7. Discussion

The paper sought to establish whether the published studies on

6.4. Quality assessment internet-delivered treatment for GAD, comparing active treatment in-
terventions with a waiting-list control, were efficacious. The meta-
Regarding the methodological quality of the studies, they can be analysis results demonstrate significant post-treatment gains on a num-
considered robust. See Fig. 2. Risk of bias graph. ber of measures for internet-delivered interventions for generalized

Fig. 2. Risk of bias graph.

 D. Richards et al. / Internet Interventions 2 (2015) 272–282 279

Table 2
Results from the meta-analysiss.

Variable Studies included Number participants treatment (control) Heterogeneity I2, Q Standardized mean difference (upper, lower) Overall effect

GADi n = 8a 321 (249) 77%: χ2 = 42.68, p b 0.00001 −0.91 (−1.25, −0.56) z = 5.10, p b .001
Worryj n = 10b, c, d, o 342 (257) 46%: χ2 = 23.90, p = 0.03 −0.74 (−0.96, −0.52) z = 6.59, p b .001
Anxietyk n = 3i 92 (76) 1%: χ2 = 3.02, p = 0.39 −0.57 (−0.86, −0.27) z = 3.81, p b .001
Depressionl n = 10e 344 (270) 66%: χ2 = 38.36, p = 0.0003 −0.63 (−0.90, −0.35) z = 4.44, p b .001
Distressm n = 4f 173 (133) 39%: χ2 = 6.54, p = 0.16 −0.91 (−1.20, −0.61) z = 6.01, p b .001
Disabilityn n = 5g 209 (152) 0%: χ2 = 4.7, p = 0.58 −0.77 (−0.97, −0.57) z = 7.63, p b .001
Quality of lifeo n = 2h 87 (70) 0%: χ2 = 1.07, p = 0.77 0.38 (0.08, 0.67) z = 2.51, p b .01
a
Andersson et al. (2012), Johansson et al. (2013), Johnston et al. (2011), Newby et al. (2013), Paxling et al. (2011), Robinson et al. (2010), Titov et al. (2009), Titov et al. (2011).
b
Andersson et al. (2012), Bell et al. (2012), Berger et al. (2013), Johnston et al. (2011), Newby et al. (2013), Paxling et al. (2011), Robinson et al. (2010), Titov et al. (2009), Titov et al.
(2010), Titov et al. (2011).
c
Berger et al. (2013), Johnston et al. (2011), Titov et al. (2010), Titov et al. (2011).
d
Johnston et al. (2011), Titov et al. (2010), Titov et al. (2011).
e
Andersson et al. (2012), Berger et al. (2013), Johannson et al. (2013), Johnston et al (2011), Newby et al. (2013), Paxling et al. (2011) ,Robinson et al. (2010), Titov et al. (2009), Titov
et al. (2011).
f
Newby et al. (2013), Robinson et al. (2010), Titov et al. (2009), Titov et al. (2011).
g
Johnston et al. (2011), Newby et al. (2013), Robinson et al. (2010), Titov et al. (2009), et al. (2011).
h
Anderson et al. (2012), Paxling et al. (2011).
i
Andersson et al. (2012), Bell et al. (2012), Paxling et al. (2011).
j
GAD-IV, Generalized Anxiety Disorder Question 7 Item Version, Generalized Anxiety Disorder Questionnaire IV.
k
Penn State Worry Questionnaire.
l
Beck Anxiety Inventory.
m
Patient Health Questionnaire 9 Item Version, Beck Depression Inventory II.
n
Kessler Distress Scale.
o
Sheehan Disability Scale, WHO Disability Assessment Schedule version 2.
o
Quality of Life Inventory.

anxiety disorder in comparison to waiting-list. Specifically, regarding studies). The evidence is limited for psychodynamic therapy, indeed
GAD symptoms a significant post-treatment effect was found for partic- one of the studies (Andersson et al., 2012) had a rather unexpected im-
ipants both across the studies included and the various interventions, provement in the waiting list.
compared to the waiting-list control participants. A similar picture is re- Although the current meta-analysis did not include a comparison
vealed for pathological worry, a central construct in generalized anxiety with face-to-face CBT studies, it is encouraging to note that outcomes
disorder, where participants in the active treatment yielded a large for pathological worry in patients with GAD could possibly be similar
post-treatment effect compared to the outcomes from the participants to what has been reported in the face-to-face CBT literature. The evi-
in the waiting-list. dence supports CBT as a highly effective treatment on symptoms of
Participants in all cases had a DSM diagnosis of GAD prior to treat- pathological worry associated with generalized anxiety disorder
ment and robust and usual post-treatment measures for GAD and path- (Covin et al., 2008). Hanrahan et al. (2013) analyzed studies that sought
ological worry were employed to assess outcomes (GAD, GAD-Q-IV and to address this primary symptom of GAD and therefore those that in-
PSWQ). With this in mind, the data from the meta-analysis supports the cluded the PSWQ as a primary outcome measure. The study of cognitive
efficacy of internet-delivered treatments for generalized anxiety disor- therapy versus a waiting-list control reported a large post-treatment ef-
der compared to waiting-list controls. The large effect size reported is fect size d = 1.81 (Hanrahan et al., 2013). In comparison, the present
similar to the post-treatment outcome from face-to-face studies of meta-analysis shows a smaller effect for pathological worry as mea-
GAD where a waiting-list control groups were also used. For instance, sured by PSWQ (d = −0.74) across the studies and for CBT-based inter-
Cuijpers et al. (2014) report a post-treatment outcome based on 38 ventions only (d = 0.87); however these effects can be considered large
comparisons (from 28 studies) of d = 0.84. (Cohen, 1988). It is important to bear in mind that the face-to-face stud-
The evidence from our analysis supports the efficacy of internet- ies are likely to have higher GAD symptom presence and lower levels of
based disorder-specific (4 studies) and transdiagnostic treatment pro- initial symptom heterogeneity; both of which Cohen's d is sensitive to.
tocols (7 studies) and those that use CBT-based treatment protocols (9 Similarly, the recent paper by Cuijpers et al. (2014), based on 20 studies

Fig. 3. Forest plot: GAD.

280 D. Richards et al. / Internet Interventions 2 (2015) 272–282

Fig. 4. Forest plot: Worry.

reported a post-treatment vs. control group effect size of d = 0.95 for post-treatment compared to a control group (Cuijpers et al., 2014). It
pathological worry. It seems that face-to-face may have a stronger im- would seem that in addition to psychotherapy (face-to-face and inter-
pact on pathological worry, although this may be confounded by the net-delivered) having a positive impact on symptoms of generalized
fact that many face-to-face interventions actually take more distressed anxiety disorder it also has a significant positive impact on comorbid de-
people so they have more space to improve. pression that may have existed in relation to primary symptoms. Addi-
tionally, impact on depression was not confounded by studies that had
mixed anxiety and depression participants.
7.1. Comorbid depression

It is not unusual to find depression as a significant comorbidity with 7.2. Psychological distress and quality of life
generalized anxiety disorder (Kessler et al., 2005b). The present study
found a significant positive shift in depression symptoms from pre- to The current study was able to analyze the results of the Kessler-10
post-treatment in comparison to waiting-list controls d = 0.63. This ef- (K-10) measure (Kessler et al., 2003) from 4 studies and demonstrated
fect is similar to that found for various psychological treatments for de- a significant improvement (d = − 0.91) in post-treatment distress in
pression in general (Cuijpers et al., 2011); for instance, an analysis of comparison to the waiting-list controls. Anxiety, generalized anxiety
215 comparisons based on 147 studies for the psychological treatment disorder, as with all mental health difficulties can cause significant dis-
for depression vs. a control group found an overall effect of d = 0.66, tress to persons and therefore realizing a significant reduction in comor-
and similar to the present analysis heterogeneity was moderate to bid distress is a positive result for internet-delivered treatments.
high. More particularly, based on 94 comparisons from 75 studies of More particularly, generalized anxiety disorder can cause serious dis-
cognitive-behavioral therapy for depression in adults vs. a control ability in one's life. The Sheehan Disability Scale (Sheehan, 1983) was
group yielded an effect size of g = 0.71 (Cuijpers et al., 2013). Interest- employed by five of the CBT-based studies in the current meta-analysis
ingly, in both of these meta-analyses when the authors adjusted for and demonstrates significant post-treatment effects (d = −0.77). The re-
publication bias, effects returned decreased to d = 0.53 and g = 0.53 re- sults on a quality of life measure used by 2 studies also confirmed a pos-
spectively. Another recent meta-analysis for psychological treatments itive and significant post-treatment effect (d = 0.38) in the current study.
for GAD, that included a small number (n = 5) of internet-delivered These findings are important given the deleterious effects that GAD can
treatments demonstrated an effect size of g = 0.71 for depression have on peoples general functioning and well-being (Wittchen, 2002).

Fig. 5. Funnel plot: GAD. Fig. 6. Funnel plot: Worry.

 D. Richards et al. / Internet Interventions 2 (2015) 272–282 281

7.3. Limitations Acknowledgments

The paper can report a number of potential limitations. First, the We would like to acknowledge with great appreciation the coopera-
number of studies included is not very high. Although in defense, meth- tion we received in collecting the discrete data for this study from the
odologically the studies are of high quality. Second, with the data, we following: Professor Gerhard Andersson, Professor Per Carlbring, Dr.
could not perform any long-term follow-up to assess maintenance of Anna Mackenzie, Dr. Jill Newby, Professor Gavin Andrews, Dr. Luke
gains post-treatment. Third, the main finding s and secondary analysis Johnston, Dr. Robert Johansson, Dr. Thomas Berger, Professor Caroline
need to be interpreted with caution as the number of study samples Bell, Dr. Frances Carter, and Dr. Johanna Boettcher. The authors would
are small, particularly, quality of life, disability, and distress. A further like to thank the peer reviewers for their fair and comprehensive review
limitation is the potential publication bias as demonstrated by the forest of this work.
plot for GAD. Although we could not complete such a comparison,
Cuijpers et al., 2014 showed that self-rated assessments in GAD has
lower effect sizes compared to clinician administered instruments. In- References
cluding studies which had more than one control group in the meta-
analysis is not ideal due to shared variance which may have affected Allen, L.B., McHugh, R.K., Barlow, D.H., 2007. Emotional Disorders a Unified Protocol Clin-
results slightly. Lastly, it is generally considered sufficient to establish ical Handbook of Psychological Disorders. 4th ed. Guilford Press, New York.
American Psychiatric Association [APA], 2013. Diagnostic and Statistical Manual of Mental
the efficacy of an intervention against a no-treatment control, but future Disorders (DSM-5). 5th ed. American Psychiatric Association, Washington D.C.
research could incorporate comparison against a realistic active control Andersson, G., Paxling, B., Roch-Norlund, P., Östman, G., Norgren, A., Almlöv, J., Silverberg,
to measure further the effects of the intervention for GAD symptoms. In F., 2012. Internet-based psychodynamic versus cognitive behavioral guided self-help
for generalized anxiety disorder: a randomized controlled trial. Psychother.
addition, waiting list controls are limited in the period of follow-up Psychosom. 81 (6), 344–355.
meaning that such analysis of the intervention may be overstated. Cau- Andrews, G., Cuijpers, P., Craske, M., McEvoy, P., Titov, N., 2010. Computer therapy for the
tion is advised regarding the results as the data is limited and highly het- anxiety and depressive disorders is effective, acceptable and practical health care: a
meta-analysis. PLoS One 5.
erogeneous, but revealing of what future research might be needed. Andrews, G., Henderson, S., Hall, W., 2001. Prevalence, comorbidity, disability and service
utilisation. Overview of the Australian National Mental Health Survey. Br. J. Psychiatry
178, 145–153.
7.4. Future research Angst, J., Vollrath, M., 1991. The natural history of anxiety disorders. Acta Psychiatr. Scand.
84, 446–452.
Barlow, D.H., 2002. Anxiety and its Disorders: The Nature and Treatment of Anxiety and
The results are encouraging for the use of internet-delivered treat-
Panic. 2nd ed. Guilford Press, New York.
ments for GAD. The pool of publish studies is still small by comparison Barlow, D.H., Ellard, K.K., Fairholme, C.P., Farchione, T.J., Boisseau, C.L., Allen, L.B.,
to, for example, studies on depression (Richards & Richardson, 2012), Ehrenreich-May, J., 2011. The Unified Protocol for Transdiagnostic Treatment of Emo-
tional Disorders: Client Workbook. Oxford University Press, New York.
consequently more research in the area would help to build the empir-
Beck, A.T., Robert, A.S., 1993. Beck AnxietyI Manual. The Psychological Corporation
ical foundations for the use of internet-delivered interventions for GAD. Harcourt Brace & Company, San Antonio, TX.
There is a greater weight of CBT-based protocols than other therapeutic Beck, A.T., Steer, R.A., Brown, G.K., 1996. Manual for the BDI-II. Psychological Corporation,
orientations and therefore future research could include greater num- San Antonio, TX.
Bell, C.J., Colhoun, H.C., Carter, F.A., Frampton, C.M., 2012. Effectiveness of computerised
bers of investigations employing varying theoretical approaches. The cognitive behaviour therapy for anxiety disorders in secondary care. Aust. N. Z.
limited number and variety of studies alongside small sample sizes J. Psychiatry 46 (7), 630–640. http://dx.doi.org/10.1177/0004867412437345.
means that there are lots of variables that could not be controlled for Berger, T., Boettcher, J., Caspar, F., 2013. Internet-based guided self-help for several anxi-
ety disorders: a randomized controlled trial comparing a tailored with a standardized
which may have a significant impact on outcomes. Future studies disorder-specific approach. Psychotherapy http://dx.doi.org/10.1037/a0032527 No
could explore these variables of internet-delivered treatments for GAD Pagination Specified..
such as the user experience, supporter function, and the possible active Bijl, R.V., Ravelli, A., van Zessen, G., 1998. Prevalence of psychiatric disorder in the general
population: results of The Netherlands Mental Health Survey and Incidence Study
ingredients including the technological tools and features regarding the (NEMESIS). Soc. Psychiatry Psychiatr. Epidemiol. 33 (12), 587–595.
presentation of content. Also as we were unable to conduct any analysis Blazer, D.G., Hughes, D., George, L.K., Swartz, M., Boyer, R., 1991. Generalized anxiety dis-
of maintenance of gains it would be important that studies include order. In: Robins, L.N., Reiger, D.A. (Eds.), Psychiatric disorders in America: The epide-
miologic catchment area studyPsychiatric Disorders in AMERICA: The Epidemiologic
follow-up and report on the results. In addition, the data did not permit
Catchment Area Study. The Free Press, New York, pp. 180–203.
us to conduct subgroups analysis for type of intervention, support type, Boettcher, J., Astrom, V., Pahlsson, D., Schenstrom, O., Andersson, G., Carlbring, P., 2014.
dropout and it is recommended that future research might consider Internet-based mindfulness treatment for anxiety disorders: a randomized controlled
trial. Behav. Ther. 45 (2), 241–253. http://dx.doi.org/10.1016/j.beth.2013.11.003.
these.
Borkovec, T.D., 1994. The nature, functions, and origins of worry. In: Davey, G.C.L., Tallis, F.
(Eds.), Worrying: Perspectives in Theory, Assessment and Treatment. John Wiley &
Sons, New York, pp. 5–33.
8. Conclusion Borkovec, T.D., Ruscio, A.M., 2001. Psychotherapy for generalized anxiety disorder. J. Clin.
Psychiatry 62 (Suppl. 11), 37–42.
The paper aimed to systematically review and analyze all published Bower, P., Gilbody, S., 2005. Stepped care in psychological therapies: access, effectiveness
and efficiency. narrative literature review. Br. J. Psychiatry 186, 11–17. http://dx.doi.
studies of internet-delivered treatments for generalized anxiety disor- org/10.1192/bjp.186.1.11.
der. Significant post-treatment gains are established for generalized Brown, T.A., O'Leary, T.A., Barlow, D.H., 2001. Generalized anxiety disorder. In: Barlow,
anxiety and symptoms of pathological worry. Results are on a par D.H. (Ed.), Clinical Handbook of Psychological Disorders, 3rd ed. Guilford Press,
New York.
with face-to-face literature regarding the efficacy of CBT for generalized Carlbring, P., Maurin, L., Törngren, C., Linna, E., Eriksson, T., Sparthan, E., Andersson, G.,
anxiety disorder. In addition, we observed significant decreases in sev- 2011. Individually-tailored, Internet-based treatment for anxiety disorders: a ran-
eral comorbid behavioral health difficulties including depression, dis- domized controlled trial. Behav. Res. Ther. 49 (1), 18–24. http://dx.doi.org/10.1016/
j.brat.2010.10.002.
tress and disability. Lastly, while the results are promising and Carter, R.M., Wittchen, H.-U., Pfister, H., Kessler, R.C., 2001. One-year prevalence of sub-
encouraging for internet-delivered interventions for generalized anxi- threshold and threshold DSM-IV generalized anxiety disorder in a nationally repre-
ety disorder further research is needed, especially to establish a more sentative sample. Depress. Anxiety 13 (2), 78–88. http://dx.doi.org/10.1002/da.1020.
Christensen, H., Batterham, P., Calear, A., 2014. Online interventions for anxiety disorders.
robust empirical foundation for their effects, to examine other theoret-
Mood Anxiety Disord. 27 (1), 7–13.
ical approaches apart from CBT, to learn more about how we can effec- Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Lawrence
tively deliver treatment, to examine follow-up for maintenance of gains, Earlbaum Associates, Hillsdale, NJ.
and explore in more detail subgroups analysis such as differences in ef- Covin, R., Ouimet, A.J., Seeds, P.M., Dozois, D.J., 2008. A meta-analysis of CBT for patholog-
ical worry among clients with GAD. J. Anxiety Disord. 22 (1), 108–116.
fects for intervention types, support types offered, and retention of Craske, M.G., Stein, M.B., Sullivan, G., et al., 2011. Disorder-specific impact of coordinated
participants. anxiety learning and management treatment for anxiety disorders in primary care.
282 D. Richards et al. / Internet Interventions 2 (2015) 272–282

Arch. Gen. Psychiatry 68 (4), 378–388. http://dx.doi.org/10.1001/archgenpsychiatry. Mewton, L., Wong, N., Andrews, G., 2012. The effectiveness of internet cognitive behav-
2011.25. ioural therapy for generalized anxiety disorder in clinical practice. Depress.Anxiety
Cuijpers, P., Andersson, G., Donker, T., van Straten, A., 2011. Psychological treatment of de- 29 (10), 843–849. http://dx.doi.org/10.1002/da.21995.
pression: results of a series of meta-analyses. Nord. J. Psychiatry 65 (6), 354–364. Meyer, T., Miller, M., Metzger, R., Borkovec, T., 1990. Development and validation of the
http://dx.doi.org/10.3109/08039488.2011.596570. Penn State Worry Questionnaire. Behav. Res. Ther. 28, 487–495.
Cuijpers, P., Berking, M., Andersson, G., Quigley, L., Kleiboer, A., Dobson, K.S., 2013. A Mohr, D.C., Ho, J., Duffecy, J., Baron, K.G., Lehman, K.A., Jin, L., Reifler, D., 2010. Perceived
meta-analysis of cognitive-behavioural therapy for adult depression, alone and in barriers to psychological treatments and their relationship to depression. J. Clin.
comparison with other treatments. Can. J. Psychiatr. 58 (7), 376–385. Psychol. 66 (4), 394–409.
Cuijpers, P., Marks, I., van Straten, A., Cavanagh, K., Gega, L., Andersson, G., 2009. Narrow, W.E., Rae, D.S., Robins, L.N., Regier, D.A., 2002. Revised prevalence based esti-
Computer-aided psychotherapy for anxiety disorders: a meta-analytic review. Cogn. mates of mental disorders in the United States: using a clinical signficance criterion
Behav. Ther. 38, 66–82. to reconcile 2 surveys' estimates. Arch. Gen. Psychiatry 59 (2), 115–123. http://dx.
Cuijpers, P., Sijbrandij, M., Koole, S.L., Huibers, M.J.H., Berking, M., Andersson, G., 2014. doi.org/10.1001/archpsyc.59.2.115.
Psychological treatment of generalized anxiety disorder: a meta-analysis. Clin. National institute for health and clinical excellence, NICE, 2011,. Generalised Anxiety Dis-
Psychol. Rev. 34 (2), 130–140. order and Panic Disorder (With or Without Agoraphobia) in Adults: Management in
Dear, B.F., Titov, N., Schwencke, G., Andrews, G., Johnston, L., Craske, M.G., McEvoy, P., 2011. Primary, Secondary and Community Care. CG113 London.
An open trial of a brief transdiagnostic internet treatment for anxiety and depression. Newby, J.M., Mackenzie, A., Williams, A.D., McIntyre, K., Watts, S., Wong, N., Andrews, G.,
Behav. Res. Ther. 49 (12), 830–837. http://dx.doi.org/10.1016/j.brat.2011.09.007. 2013. Internet cognitive behavioural therapy for mixed anxiety and depression: a
Draper, M., Rees, C.S., Nathan, P.R., 2008. Internet-based self-management of generalised randomized controlled trial and evidence of effectiveness in primary care. Psychol.
anxiety disorder: a preliminary study. Behav. Chang. 25 (04), 229–244. http://dx.doi. Med. 43 (12), 2635–2648. http://dx.doi.org/10.1017/S0033291713000111.
org/10.1375/bech.25.4.229. Newman, M.G., Szkodny, L.E., Llera, S.J., Przeworski, A., 2011. A review of technology-
Dugas, M.J., Robichaud, M., 2007. Cognitive-Behavioral Treatment for Generalized Anxiety assisted self-help and minimal contact therapies for anxiety and depression: is
Disorder. Routledge, New York. human contact necessary for therapeutic efficacy? Clin. Psychol. Rev. 31 (1),
Faravelli, C., Guerrini Degl'Innocenti, B., Giardinelli, L., 1989. Epidemiology of anxiety dis- 89–103. http://dx.doi.org/10.1016/j.cpr.2010.09.008.
orders in Florence. Acta Psychiatr. Scand. 79, 308–312. Newman, M.G., Zuellig, A.R., Kachin, K.E., Constantino, M.J., Przeworski, A., Erickson, T.,
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 1997. Structured Clinical Interview for Cashmnan-McGrath, L., 2002. Preliminary reliability and validity of the Generalized
DSM-IV-Clinical Version (SCID-CV). American Psychiatric Press, Inc., Washington, D.C. Anxiety Disorder Questionnaire-IV: a revised self-report diagnostic measure of gen-
Grant, B.F., Hasin, D.S., Stinson, F.S., Dawson, D.A., June Ruan, W., Goldstein, R.B., Huang, B., eralized anxiety disorder. Behav. Ther. 33, 215–233.
2005. Prevalence, correlates, co-morbidity, and comparative disability of DSM-IV Paxling, B., Almlov, J., Dahlin, M., Carlbring, P., Breitholtz, E., Eriksson, T., Andersson, G.,
generalized anxiety disorder in the USA: results from the National Epidemiologic Sur- 2011. Guided internet-delivered cognitive behavior therapy for generalized anxiety
vey on Alcohol and Related Conditions. Psychol. Med. 35, 1747–1759. disorder: a randomized controlled trial. Cogn. Behav. Ther. 40, 159–173.
Hanrahan, F., Fielda, A.P., Jones, F.W., Davey, G.C.L., 2013. A meta-analysis of cognitive Przeworski, A., Newman, M.G., 2006. Efficacy and utility of computer-assisted cognitive
therapy for worry in generalized anxiety disorder. Clin. Psychol. Rev. 33 (1), 120–132. behavioural therapy for anxiety disorders. Clin. Psychol. 10 (2), 43–53. http://dx.
Higgins, J.P.T., Thompson, S.G., Deeks, J.J., Altman, D.G., 2003. Measuring inconsistency in doi.org/10.1080/13284200500378779.
meta-analyses. BMJ 327 (7414), 557–560. http://dx.doi.org/10.1136/bmj.327.7414.557. Reger, M., Gahm, G., 2009. A meta-analysis of the effects of internet-and computer-based
Higgins, J.P.T., Green, S. (Eds.), 2009. Cochrane Handbook for Systematic Reviews of Inter- cognitive-behavioral treatments for anxiety. J. Clin. Psychol. 65, 53–75.
ventions. John Wiley & Sons Ltd, Sussex, England. Richards, D., Richardson, T., 2012. Computer-based psychological treatments for depres-
Hunot, V., Churchill, R., Teixeira, V., Silva de Lima, M., 2007. Psychological therapies for sion: a systematic review and meta-analysis. Clin. Psychol. Rev. 32 (4), 329–342.
generalised anxiety disorder. Cochrane Database Syst. Rev. http://dx.doi.org/10. http://dx.doi.org/10.1016/j.cpr.2012.02.004.
1002/14651858.CD001848.pub4. Robinson, E., Titov, N., Andrews, G., McIntyre, K., Schwencke, G., Solley, K., 2010. Internet
Hunt, C., Issakidis, C., Andrews, G., 2002. DSM-IV generalized anxiety disorder in the treatment for generalized anxiety disorder: a randomized controlled trial comparing
Australian National Survey of Mental Health and Well-Being. Psychol. Med. 32, clinician vs. technician assistance. PLoS ONE 5 (6), e10942. http://dx.doi.org/10.1371/
649–659. journal.pone.0010942.
Jenkins, R., Lewis, G., Bebbington, P., Brugha, T., Farrell, M., Gill, B., Meltzer, H., 1997. The Roy-Byrne, P.P., Wagner, A., 2004. Primary care perspectives on generalized anxiety disor-
National Psychiatric Morbidity Surveys of Great Britain–initial findings from the der. J. Clin. Psychiatry 65 (Suppl. 13), 20–26.
household survey. Psychol. Med. 27, 775–789. Sanderson, W.C., Wetzler, S., Beck, A.T., Betz, F., 1994. Prevalence of personality disorders
Johansson, R., Björklund, M., Hornborg, C., Karlsson, S., Hesser, H., Ljótsson, B., Andersson, among patients with anxiety disorders. Psychiatry Res. 51, 167–174.
G., 2013. Affect-focused psychodynamic psychotherapy for depression and anxiety Sheehan, D., 1983. The Anxiety Disease. Scribner's, New York.
through the Internet: a randomized controlled trial. PeerJ 1, e102. http://dx.doi.org/ Sheehan, D., Lecrubier, Y., Sheehan, K., Amorim, P., Janavs, J., Weiller, E., Dunbar, G., 1998.
10.7717/peerj.102. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and
Johnston, L., Titov, N., Andrews, G., Spence, J., Dear, B.F., 2011. A RCT of a transdiagnostic validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10.
Internet-delivered treatment for three anxiety disorders: examination of support J. Clin. Psychiatry 59 (Suppl. 20), 22–33.
roles and disorder-specific outcomes. PLoS One 6 (11). http://dx.doi.org/10.1371/ Spitzer, R.L., Kroenke, K., Williams, J., 1999. Validation and utility of a self-report version of
journal.pone.0028079. PRIME-MD: the PHQ primary care study. J. Am. Med. Assoc. 282, 1737–1744.
Kessler, R.C., Berglund, P., Demler, O., Jin, R., Merikangas, K.R., Walters, E.E., 2005a. Life- Spitzer, R.L., Kroenke, K., Williams, J.W., Löwe, B., 2006. A brief measure for assessing gen-
time prevalence and age-of-onset distributions of DSM-IV disorders in the National eralized anxiety disorder: the GAD-7. Arch. Intern. Med. 166 (10), 1092–1097. http://
Comorbidity Survey Replication. Arch. Gen. Psychiatry 62 (6), 593–602. http://dx. dx.doi.org/10.1001/archinte.166.10.1092.
doi.org/10.1001/archpsyc.62.6.593. Stein, M.B., Heimberg, R.G., 2004. Well-being and life satisfaction in generalized anxiety
Kessler, R.C., Chiu, W., Demler, O., Walters, E.E., 2005b. Prevalence, severity, and comor- disorder: comparison to major depressive disorder in a community sample.
bidity of 12-month DSM-IV disorders in the national comorbidity survey replication. J. Affect. Disord. 79 (1), 161–166.
Arch. Gen. Psychiatry 62 (6), 617–627. http://dx.doi.org/10.1001/archpsyc.62.6.617. Sunderland, M., Wong, N., Hilvert-Bruce, Z., Andrews, G., 2012. Investigating trajectories
Kessler, R.C., Barker, P.R., Colpe, L.J., Epstein, J.F., Gfroerer, J.C., Hiripi, E., Zaslavsky, A.M., of change in psychological distress amongst patients with depression and generalised
2003. Screening for serious mental illness in the general population. Arch. Gen. Psy- anxiety disorder treated with internet cognitive behavioural therapy. Behav. Res.
chiatry 60 (2), 184–189. Ther. 50 (6), 374–380. http://dx.doi.org/10.1016/j.brat.2012.03.005.
Kessler, R.C., Berglund, P., Chiu, W.T., Demler, O., Heeringa, S., Hiripi, E., Zheng, H., 2004. Titov, N., Andrews, G., Johnston, L., Robinson, E., Spence, J., 2010. Transdiagnostic Internet
The US National Comorbidity Survey Replication (NCS-R): design and field proce- treatment for anxiety disorders: a randomized controlled trial. Behav. Res. Ther. 48
dures. Int. J. Methods Psychiatr. Res. 13 (2), 69–92. (9), 890–899. http://dx.doi.org/10.1016/j.brat.2010.05.014.
Klein, B., Meyer, D., Austin, D.W., Kyrios, M., 2011. Anxiety online: a virtual clinic: prelim- Titov, N., Andrews, G., Robinson, E., Schwencke, G., Johnston, L., Solley, K., Choi, I., 2009.
inary outcomes following completion of five fully automated treatment programs for Clinician-assisted internet-based treatment is effective for generalized anxiety disor-
anxiety disorders and symptoms. J. Med. Internet Res 13 (4), e89. http://dx.doi.org/ der: randomised controlled trial. Aust. N. Z. J. Psychiatry 43, 905–912.
10.2196/jmir.1918. Titov, N., Dear, B.F., Schwencke, G., Andrews, G., Johnston, L., Craske, M.G., McEvoy, P.,
Kohn, R., Saxena, S., Levav, I., Saraceno, B., 2004. The treatment gap in mental health care. 2011. Transdiagnostic internet treatment for anxiety and depression: a randomised
Bull. World Health Organ. 82, 858–866. controlled trial. Behav. Res. Ther. 49 (8), 441–452. http://dx.doi.org/10.1016/j.brat.
Kroenke, K., Spitzer, R.L., Williams, J.B., Monahan, P.O., Lowe, B., 2007. Anxiety disorders in 2011.03.007.
primary care: prevalence, impairment, comorbidity, and detection. Ann. Intern. Med. Wells, G., 1999. A metacognitive model and therapy for generalized anxiety disorder. Clin.
146 (5), 317–325. Psychol. Psychother. 6 (2), 86–95.
Lepine, J.P., 2002. The epidemiology of anxiety disorders: prevalence and societal costs. Wittchen, H.U., 2002. Generalized anxiety disorder: prevalence, burden, and cost to soci-
J. Clin. Psychiatry 63 (Suppl. 14), 4–8. ety. Depress.Anxiety 16 (4), 162–171. http://dx.doi.org/10.1002/da.10065.
Loebach Wetherell, J., Thorp, S.R., Patterson, T.L., Golshan, S., Jeste, D.V., Gatz, M., 2004. Zou, J.B., Dear, B.F., Titov, N., Lorian, C.N., Johnston, L., Spence, J., Sachdev, P., 2012. Brief
Quality of life in geriatric generalized anxiety disorder: a preliminary investigation. internet-delivered cognitive behavioral therapy for anxiety in older adults: a feasibil-
J. Psychiatr. Res. 38 (3), 305–312. ity trial. J. Anxiety Disord. 26 (6), 650–655. http://dx.doi.org/10.1016/j.janxdis.2012.
Massion, A.O., Warshaw, M.G., Keller, M.B., 1993. Quality of life and psychiatric morbidity in 04.002.
panic disorder and generalized anxiety disorder. Am. J. Psychiatry 150 (4), 600–607.
Mayo-Wilson, E., Montgomery, P., 2013. Media-Delivered Cognitive Behavioural Therapy
and Behavior (Self-help) for Anxiety Disorders in Adults. Cochrane Collaboration.
Cochrane Collaboration, New Jersey, USA.