Effectiveness of Transdiagnostic Cognitive Behaviour Therapy for Anxiety and Depression in Adults: a systematic Review and Meta-analysis



Download 103.66 Kb.
Page1/4
Date20.01.2019
Size103.66 Kb.
  1   2   3   4
left running head: P. Andersen et al.

right running head: Effectiveness of transdiagnostic CBT


Effectiveness of Transdiagnostic Cognitive Behaviour Therapy for Anxiety and Depression in Adults: A Systematic Review and Meta-analysis
Philip Andersen, Paul Toner and Martin Bland

University of York, UK
Dean McMillan

University of York, and Hull York Medical School, UK

Correspondence to Dean McMillan, Hull York Medical School and Department of Health Sciences, Room ARC/205, University of York, York YO10 5DD, UK. E-mail: dean.mcmillan@york.ac.uk



Background: Transdiagnostic Cognitive Behaviour Therapy (CBT) seeks to identify core cognitive-behavioural processes hypothesized to be important across a range of disorders and to develop a treatment that targets these. This contrasts with standard CBT approaches that are disorder-specific. Proponents of transdiagnostic CBT suggest that it may offer advantages over disorder-specific CBT, but little is known about the effectiveness of this approach. Aims: The review aimed to summarize trial-based clinical and cost-effectiveness data on transdiagnostic CBT for anxiety and depression. Method: A systematic review of electronic databases, including peer-reviewed and grey literature sources, was conducted (n = 1167 unique citations). Results: Eight trials were eligible for inclusion in the review. There was evidence of an effect for transdiagnostic CBT when compared to a control condition. There were no differences between transdiagnostic CBT and active treatments in two studies. We found no evidence of cost-effectiveness data. Conclusions: Quality assessment of the primary studies indicated a number of methodological concerns that may serve to inflate the observed effects of transdiagnostic approaches. Although there are positive signs of the value of transdiagnostic CBT, there is as yet insufficient evidence to recommend its use in place of disorder-specific CBT.
Keywords: Transdiagnostic, cognitive behaviour therapy, systematic review, meta-analysis, methodological quality, evidence-base

Introduction

Depression and anxiety are highly prevalent (Kessler, Chiu, Demler, Merikangas and Walters, 2005; McManus, Meltzer, Brugha, Bebbington and Jenkins, 2009), have substantial impacts on quality of life and functioning for the individual (Haslam, Atkinson, Brown and Haslam, 2005; Paul and Moser, 2009) and are associated with substantial economic costs (Das-Munshi et al., 2008; Health and Safety Executive, 2012). Cognitive behaviour therapy (CBT) is an effective treatment for a range of depressive (Butler, Chapman, Forman and Beck, 2006; Hofmann, Asnaani, Vonk, Sawyer and Fang, 2012) and anxiety conditions (Olatunji, Cisler and Deacon, 2010; Otte, 2011), but the vast majority of evidence for the effectiveness of this approach is based on highly differentiated treatment protocols for specific diagnoses (NICE, 2011). An alternative approach, termed transdiagnostic CBT (tCBT), aims to identify a small number of cognitive and behavioural processes that are common across a range of depressive and anxiety conditions. These core processes are then used to develop a single treatment that can be applied across a range of presentations.

Proponents of tCBT have put forward a number of potential advantages to this approach. These include the possibility of offering more effective, efficient treatment of co-morbid presentations (Borkovec, Abel and Newman, 1995), something that is common in depression and anxiety presentations (Kessler et al., 2005). Furthermore, tCBT may reduce the training needs of therapists (Schmidt et al., 2012). Since tCBT by definition should be effective for a variety of different presentations, formation of ideally sized groups for group-delivered CBT should be simplified (Norton and Philipp, 2008).

There are potential disadvantages to tCBT, not as widely discussed in the literature. Clinical effectiveness may be diluted compared to diagnosis-specific treatments, because treatments are designed to treat a wide variety of disorders (Craske et al., 2007). It is unclear whether tCBT treatment is as acceptable to clients as disorder-specific CBT. When delivered in groups with a mixture of presentations, it may be difficult for clients to empathize with and learn as much from each other compared to groups in which clients have similar presentations (McEvoy, Nathan and Norton, 2009).

Three previous reviews have examined the effectiveness of tCBT. The first, Norton and Philipp (2008), was limited to anxiety studies. Of the nine studies reviewed, two were randomized controlled trials (RCTs) (Erickson, Janeck and Tallman, 2007; Schmidt et al., 2012) and one used a quasi-randomized design (Norton and Hope, 2005); the remainder used uncontrolled designs. The review calculated a large overall within-group effect size of transdiagnostic anxiety treatments (d = 1.29, 95% CI 0.66 to 1.93), compared to a small within-group effect size for control conditions (d = 0.14, 95% CI −0.21 to 0.49). The authors concluded that tCBT had considerable clinical utility, the preliminary data supporting the efficacy of the interventions.

The second review, McEvoy et al. (2009), included 10 studies, with one RCT (Erickson et al., 2007). The same quasi-randomized study was included, cited as two separate studies (Norton, Hayes and Hope, 2004; Norton and Hope, 2005). The remainder were uncontrolled studies. No meta-analysis of results was performed but a narrative synthesis concluded that tCBT protocols were highly promising, although additional research was required.

The third review, Reinholt and Krogh (2014), included five RCTs (Erickson, Janeck and Tallman, 2007; Farchione et al., 2012; Norton and Barrera, 2012; Roy-Byrne et al., 2010; and Schmidt et al., 2012), the quasi-randomized Norton and Hope (2005) and six uncontrolled studies. They reported a moderate combined effect size for all studies (SMD = -0.681, 95% CI -0.903 to -0.458) favouring tCBT. The authors were more cautious in their conclusions than the previous reviewers.

Although the reviews were broadly positive about the effectiveness of tCBT approaches, caution is needed before accepting these conclusions. The reviews identified only a small number of RCTs; the majority of studies were based on uncontrolled designs. The RCTs all compared tCBT to non-active control conditions, with the exception of Norton and Barrera (2012). There were a number of methodological limitations of the reviews when evaluated against standard guidelines for the conduct of systematic reviews (Moher, Liberati, Tetzlaff, Altman and The PRISMA Group, 2009). There was limited description of search strategies, and two reviews were limited to peer-reviewed publications, which may lead to an overestimate of effects because of publication bias (Dwan et al., 2008; Song, Eastwood, Gilbody, Duley and Sutton, 2000). There was limited evaluation of the methodological quality of included studies. Two reviews were conducted by researchers involved in tCBT protocol development. There is evidence that researcher allegiance can lead to inflated effect estimates in RCTs of psychological treatments (Luborsky et al., 1999). The case for tCBT would be strengthened by an additional independent review of the current evidence base.


Method

We adhered to the Centre for Reviews and Dissemination (2009) guidelines in the conduct of the review and PRISMA guidelines (Moher et al., 2009) in the reporting of the review.


Search strategy

The following databases were searched to cover peer-review and grey literature sources: PsycINFO, Medline, Embase, Cochrane Library, ASSIA, Web of Knowledge, OAIster, Open Grey, Trip Database and ZETOC. Additional searches were conducted in NHS EED, HTA Database, CEA Registry and RePEc health economic databases to identify relevant cost-effectiveness data. Databases were searched from inception to June 2013, with no publication status or language restrictions imposed.

Reference lists of included studies were examined, reverse-citation searches of included studies were conducted and websites of researchers and study groups in the field of tCBT were accessed to identify additional studies.


Search terms

Search terms, including free text and thesauri terms, were developed for PsycINFO and adapted for other databases (Appendix A). Terms covered two broad constructs, CBT and transdiagnostic treatment, combined using the Boolean AND. Search terms for CBT were adapted, with permission, from a systematic review of low-intensity psychological interventions (Rodgers et al., 2012). Search terms designed to capture the “transdiagnostic” construct were derived from searches of books, studies and articles uncovered during scoping searches. Thesauri available within the Ovid SP database were used to further inform search terms related to CBT.


Study inclusion and exclusion criteria

A priori inclusion and exclusion criteria were developed around standard criteria.



Participants. Adults (aged 16 years or over). Participants at baseline met diagnostic criteria for an anxiety or depressive disorder established by gold standard structured clinical interviews or scored above a clinical cut-off point on a standardised severity measure. To ensure treatment was transdiagnostic, participants within studies must not have had uniform diagnoses.

Intervention. The intervention had to be described as tCBT (or use related terms such as Unified Protocol, mixed-diagnosis CBT, broad-spectrum CBT).

Comparator. Any comparison condition, including control or other active treatments.

Outcomes. Depression, anxiety or generic psychological wellbeing measured as a severity score or a dichotomous outcome (e.g. depressed, not depressed) and health economic data.

Study design. Randomized controlled trials. Quasi-randomized and uncontrolled trials were excluded.
Selection of studies

Two reviewers (PA, PT) used a pre-piloted eligibility form to assess studies. Disagreements were resolved through consensus and, where necessary, discussion with a third reviewer (DMcM). Titles and abstracts were first examined, and full papers obtained for studies passing this initial sift. Full papers were then re-examined using the eligibility form to determine final inclusion.


Data extraction

Data were extracted to a pre-piloted data extraction form by the primary reviewer (PA) and checked by the secondary reviewer (PT). Extracted data included study name, year and type of publication, authors, study location and setting, design, study sample characteristics, inclusion and exclusion criteria, description of interventions and controls, follow-up length, outcomes reported and data necessary to calculate effect sizes (e.g. means, standard deviations, sample size).


Quality assessment

Methodological quality and sources of bias were assessed using the Cochrane risk of bias assessment tool (Higgins et al., 2011) and an additional quality tool specifically tailored to assess the quality of psychological RCTs (Yates, Morley, Eccleston and Williams, 2005).


Data synthesis

Standardized mean differences (SMD) with 95 percent confidence intervals using Hedges’s adjusted g (Hedges, 1981) were calculated for anxiety, depression and general psychological wellbeing severity measures.



Pre-planned comparisons. Results were grouped first by intervention type (individual or group), second by comparator (control condition or active treatment) and finally by outcome measure (anxiety, depression, general psychological wellbeing). If two or more studies were similar in terms of the intervention, comparator and outcome type, a meta-analysis was conducted. Heterogeneity was assessed using the I2 statistic, with values over 50% taken to indicate substantial heterogeneity. Analyses were conducted in Review Manager 5 (Cochrane Collaboration, 2012).
Results


Electronic database searches identified 2306 citations. Seven additional unique records were identified through other sources. After removing duplicates, 1167 unique citations remained. Screening titles and abstracts excluded 1133 citations, leaving 34 full-text articles to be assessed. Of these, eight were judged eligible for inclusion in the review. Four studies were meta-analysed (see Figure 1).

Figure 1. Search results and study selection flowchart
Of 34 full-text articles assessed, 26 were excluded for the following reasons: 11 studies did not use randomization or used quasi-random methods of assigning participants to treatment; nine studies were excluded for not using recognizable tCBT treatment protocols; participants in six studies did not meet diagnostic inclusion criteria (Appendix B).

Table 1. Characteristics of included studies

Study

Setting and sample

Intervention/s

Comparator

Treatment focus/outcomes

Erickson et al. (2007)

Setting: Teaching hospitals, University training clinic, Canada

Age (years): (Mean±SD)

Intervention: 40.7±11.8

Comparator: 41.0±11.1

Overall: 40.9±11.4*

Per cent female: 63.8

Ethnicity: Not reported


Transdiagnostic CBT

11 X 120 minutes weekly groups

Senior doctoral-level psychologist with senior graduate student

N = 73 (47 at post-treatment)



Waitlist / delayed treatment

N = 79 (41 at post-treatment)

Note: Waitlist group began treatment 1 week after CBT groups completed treatment


Treatment focus: Anxiety

Diagnosis: SCID

Anxiety severity: BAI

Depression severity: N/A

Generic measure: N/A

Measurement points: Baseline, post-treatment, 6-months post-treatment



Farchione et al. (2012)

Setting: University study clinic, USA

Age (years): (Mean±SD)

Intervention: 29.38±9.86

Comparator: 30.64±9.15

Overall: 29.75±9.66*

Per cent female: 59.5

Ethnicity: 94.6% Caucasian


Transdiagnostic CBT – Unified Protocol (UP)

18 X 60 minutes individual sessions

Doctoral students, 2-4 years’ experience; licenced doctoral-level psychologist, 7 years’ experience

N = 26 (22 at post-treatment)



Waitlist / delayed treatment

N = 11 (10 at post-treatment)

Note: Waitlist group began treatment 16 weeks after start of trial


Treatment focus: Anxiety

Diagnosis: ADIS-IV

Anxiety severity: HARS/BAI

Depression severity: HRSD/BDI-II

Generic measure: N/A

Measurement points: Baseline, post-treatment, 6-months post-treatment



Johnston et al. (2011)

Setting: Internet, Australia

Age (years): (Mean±SD)

CL Group: 43.74±13.36

CO Group: 38.63±11.56

Comparator: 42.36±13.20

Overall: 41.62±12.83

Per cent female: 58.8

Ethnicity: Not reported



1. Clinician assisted (CL) transdiagnostic iCBT (Anxiety Program)

8 units over 10 weeks

Clinical psychologist

N = 47 (42 at post-treatment)

2.Coach assisted (CO) transdiagnostic iCBT (Anxiety Program)

8 units over 10 weeks

Registered psychologist

N = 46 (39 at post-treatment)



Waitlist / delayed treatment

N = 46 (41 at post-treatment)

Note: Waitlist group began treatment immediately after iCBT groups completed treatment


Treatment focus: Anxiety

Diagnosis: MINI

Anxiety severity: GAD-7

Depression severity: PHQ-9

Generic measure: DASS-21

Measurement points: Baseline, post-treatment, 3-months post-treatment



Norton (2012)

Setting: University anxiety clinic, USA

Age (years): (Mean±SD)

Overall: 32.98±10.73

Per cent female: 62.1

Ethnicity: 58.6% Caucasian

21.8% Hispanic/Latino

9.2% African American

4.6% Asian American

5.7% other or mixed


Transdiagnostic CBT

12 x 120 minutes weekly groups

Doctoral-level graduate students with experienced senior graduate co-therapists

N = 65 (37 at post-treatment)



Relaxation

12 x 120 mins weekly groups

Doctoral-level graduate students with experienced senior graduate co-therapists

N = 22 (7 at post-treatment)



Treatment focus: Anxiety

Diagnosis: ADIS-IV

Anxiety severity: BAI

Depression severity: N/A

Generic measure: CGI-S

Measurement points: Baseline, mid-treatment, post-treatment



Norton & Barrera (2012)

Setting: University anxiety clinic, USA

Age (years): (Mean±SD)

Overall: 31.46±8.93

Per cent female: 50.0

Ethnicity: 54.3% Caucasian

23.9% Hispanic/Latino

10.9% African American

6.5% Asian American

4.3% Other or mixed


Transdiagnostic CBT

12 x 120 minutes weekly groups

Doctoral-level graduate students with experienced senior graduate co-therapists

N = 23 (16 at post-treatment)

(57 total randomized, 12 declined treatment – allocation not stated)


Diagnosis-specific group CBT

12 x 120 mins weekly groups

Doctoral-level graduate students with experienced senior graduate co-therapists

N = 23 (12 at post-treatment)

(See note in ‘Intervention/s’ column)


Treatment focus: Anxiety

Diagnosis: ADIS-IV

Anxiety severity: STAI

Depression severity: BDI-II

Generic measure: CGI-S

Measurement points: Baseline, mid-treatment, post-treatment



Schmidt et al. (2012)

Setting: University outpatient clinic, USA

Age (years): (Mean±SD)

Intervention: 37.5±11.3

Comparator: 34.6±9.9

Overall: 36.3±10.7*

Per cent female: 71.9

Ethnicity: 83% Caucasian

10% African American

7% other


Transdiagnostic CBT – False Safety Behavior Elimination Therapy (F-SET)

10 x 120 minutes weekly groups

Master's-level therapists (1 to 2 years clinical experience), experienced Ph.D. postdoctoral fellow

N = 57 (53 at post-treatment)



Waitlist / delayed treatment

N = 39 (39 at post-treatment)

Note: Waitlist group began treatment immediately after CBT groups completed treatment


Treatment focus: Anxiety

Diagnosis: SCID

Anxiety severity: SPRAS

Depression severity: BDI-II

Generic measure: CGI-S

Measurement points: Baseline, post-treatment,

6-months post-treatment (F-SET group only)


Titov et al. (2010)

Setting: Internet, Australia

Age (years): (Mean±SD)

Intervention: 38.6±12.0

Comparator: 40.5±14.1

Overall: 39.5±13.0

Per cent female: 67.9

Ethnicity: Not reported


Transdiagnostic iCBT (Anxiety Program)

6 units over 8 weeks

Clinical psychologist via weekly telephone calls or instant messaging

N = 42 (36 at post-treatment)



Waitlist / delayed treatment

N = 44 (36 at post-treatment)

Note: Waitlist group began treatment immediately after iCBT group completed treatment


Treatment focus: Anxiety

Diagnosis: MINI

Anxiety severity: GAD-7

Depression severity: PHQ-9

Generic measure: DASS-21

Measurement points: Baseline, post-treatment, 3-months post-treatment (treatment group only)



Titov et al. (2011)

Setting: Internet, Australia

Age (years): (Mean±SD)

Intervention: 44.8±14.9

Comparator: 42.9±14.5

Overall: 43.9±14.6

Per cent female: 73.0

Ethnicity: Not reported


Transdiagnostic iCBT (Wellbeing Program)

8 units over 10 weeks

Clinical psychologist via weekly telephone calls or instant messaging

N = 39 (34 at post-treatment)



Waitlist / delayed treatment

N = 38 (35 at post-treatment)

Note: Waitlist group began treatment immediately after iCBT group completed treatment


Treatment focus: Anxiety or depression

Diagnosis: MINI

Anxiety severity: GAD-7

Depression severity: PHQ-9

Generic measure: DASS-21

Measurement points: Baseline, post-treatment, 3-months post-treatment (treatment group only)



Notes: * Calculated using pooled variance; ADIS-IV = Anxiety Disorders Interview Schedule for the DSM-IV; BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory (1996 revision); CBT = cognitive behavior therapy; CGI-S = Clinical Global Impressions - Severity scale; DASS-21 = Depression Anxiety Stress Scales - short-form version; F-SET = False Safety Behavior Elimination Therapy; GAD-7 = Generalized Anxiety Disorder 7-item scale; iCBT = internet CBT; MINI = Mini International Neuropsychiatric Interview Version 5.0.0; N = number of participants; N/A = not available; PHQ-9 = Patient Health Questionnaire - 9; SCID = Structured Clinical Interview for DSM-IV Axis I Disorders; SD = standard deviation; HARS = Hamilton Anxiety Rating Scale; HRSD = Hamilton Rating Scale for Depression; SPRAS = Sheehan Patient-Rated Anxiety Scale; STAI = State-Trait Anxiety Inventory
Description of included studies

Table 1 summarizes the descriptive characteristics of the eight studies that met inclusion criteria for the review (Erickson et al., 2007; Farchione et al., 2012; Johnston, Titov, Andrews, Spence and Dear, 2011; Norton, 2012; Norton and Barrera, 2012; Schmidt et al., 2012; Titov, Andrews, Johnston, Robinson and Spence, 2010; Titov et al., 2011).



Table 2 summarizes the principal diagnosis of the total sample. Social phobia (32.3%), panic disorder (25.2%) and generalized anxiety disorder (29%) comprised the overwhelming majority of diagnoses. Depression was the principal diagnosis in only 5.2% of the entire sample.
Table 2. Principal diagnosis of included participants

Principal diagnosis

N

%

Social phobia

226

30.9

Panic disorder, with or without agoraphobia

177

24.2

Generalized anxiety disorder

203

27.7

Posttraumatic stress disorder

17

2.3

Obsessive-compulsive disorder

25

3.4

Specific phobia

8

1.1

Anxiety disorder not otherwise specified

4

0.5

Social phobia and anxiety NOS

1

0.1

Generalized anxiety disorder and social phobia

1

0.1

Obsessive-compulsive disorder and panic disorder

1

0.1

Major depressive disorder

38

5.2

Not reported

31

4.2

Total

732

100



Quality assessment of included studies

Table 3 summarizes the quality assessment of included studies using the Cochrane risk of bias tool. All studies were deemed at high risk of bias for blinding of participants and personnel, because this is typically inevitable in a trial of a psychological intervention. All studies were judged at high risk of bias in “other sources of bias” because of potential for research allegiance effects (treatment developers were the researchers evaluating their effectiveness). Unclear random sequence generation prompted an unknown risk of bias for this criterion in four studies (Erickson et al., 2007; Farchione et al., 2012; Norton, 2012; Norton and Barrera, 2012). Erickson et al. (2007) had an unknown risk of bias because of unclear information concerning blinding of assessors. This study was also assessed to be at high risk of bias due to selective reporting; two further were rated as having an unknown risk of bias on this quality item (Norton, 2012; Norton and Barrera, 2012).


Table 3. Cochrane risk of bias assessment tool summary




Random sequence generation

Allocation concealment

Blinding of participants and personnel

Blinding of outcome assessment

Incomplete outcome data

Selective reporting

Other bias

Erickson 2007

?

?



?







Farchione 2012

?

?



+

+

+



Johnston 2011

+

+



+

+

+



Norton 2012a

?

?



+



?



Norton 2012b

?

?



+



?



Schmidt 2012

+

+





+

+



Titov 2010

+

+



+

+

+



Titov 2011

+

+



+

+

+



Notes: '+' low risk of bias; '−' high risk of bias; '?' unknown risk of bias
Assessment of treatment quality within included studies

The Yates assessment tool (Yates et al., 2005) was used to assess quality of treatment administered in source studies. Items assessing description of treatment content and setting, description of treatment duration and manualization of treatment protocols were rated as present in all source studies. One study did not assess clinician adherence to the manual (Farchione et al., 2012). One study provided no information concerning therapist training (Erickson et al., 2007). Evidence of a lack of client engagement was identified in two studies (Norton, 2012; Norton and Barrera, 2012) (Table 4).




Share with your friends:
  1   2   3   4


The database is protected by copyright ©sckool.org 2019
send message

    Main page