Home > DARE Reviews > Is computerised CBT really helpful for...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews [Internet]. York (UK): Centre for Reviews and Dissemination (UK); 1995-.

Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews [Internet].

Is computerised CBT really helpful for adult depression? A meta-analytic re-evaluation of CCBT for adult depression in terms of clinical implementation and methodological validity

M So, S Yamaguchi, S Hashimoto, M Sado, TA Furukawa, and P McCrone.

Review published: 2013.

CRD summary

This review concluded that computerised cognitive behavioural therapy led to significant reduction of adult depression symptoms at post-treatment but was not maintained at long-term follow-up. No significant improvement in function was observed at post-intervention and dropout was high. These conclusions appear reliable; more recent studies may have become available since this review's search in 2011.

Authors' objectives

To evaluate the short- and long-term effectiveness, functional improvement and dropout associated with computerised cognitive behavioural therapy (CCBT) for adult depression.


Five databases (including MEDLINE, EMBASE and Cochrane CENTRAL) were searched to July 2011. Search terms were reported in an appendix. An online source (www.controlledtrials.com) was searched.

Study selection

Randomised controlled trials (RCTs) that evaluated the effects of guided/unguided CCBT in adults (18 years or over) with depression were eligible for inclusion. The primary outcome was depression, measured at intervention end point and (if applicable) long-term follow-up (>6 months). Secondary outcomes were post-intervention function and dropout. Studies had to have used reliable and standardised rating-scales at baseline and follow-up, report proper allocation concealment and (at the very least) single blinding of outcome assessors. Studies that used medication or other psychotherapies were included. Studies on in-patients and patients with comorbidities (such as psychotic disorders, manic status, dementia, severe physical conditions) were excluded.

The included studies were conducted in Sweden, Australia, USA, the Netherlands and Germany. Most recruited patients from the community. The mean age of patients ranged from 22.6 to 55 years. All of the studies provided internet-based CBT and the number of modules ranged from four to nine. Half of the studies delivered guided CCBT; the rest delivered unguided, technician-guided or clinician-guided CCBT. Most control groups received treatment as usual (not defined) or were put on a wait list. Outcome and function measures varied.

Multiple reviewers selected the studies for inclusion in the review.

Assessment of study quality

No details of any quality assessment process were reported, although the inclusion criteria specified that studies had to report adequate allocation, concealment, blinding of outcome assessors and total dropout rates (criteria not reported).

Data extraction

Data on the outcomes were extracted to calculate standardised mean differences or relative risks, with corresponding 95% confidence intervals.

The authors did not state how many reviewers extracted data.

Methods of synthesis

Effect estimates and 95% confidence intervals were pooled using random-effects models. Statistical heterogeneity was assessed using the Cochran's Q and the Ι² statistic. Subgroup analyses were performed according to the type of control (wait list, treatment as usual). Sensitivity analyses were performed by excluding studies that used the Beck Depression Inventory I or II and by excluding studies according to differences in attrition rates and imputation techniques. Where studies used a number of depression outcome measures, the main outcome measure (the primary endpoint/the endpoint first reported) was used. Publication bias was assessed using a funnel plot, Begg's and Egger's tests, and the trim-and-fill method.

Results of the review

Fourteen RCTs reporting on 16 comparisons were included in the review (3,286 patients, range 45 to 525 patients per comparison). Length of follow-up ranged from two weeks to 18 months.

Compared with controls, depression symptoms in CCBT patients were statistically significantly reduced at intervention endpoint (SMD -0.48, 95% CI -0.63 to -0.33; 16 comparisons; Ι²=71%). Similar results were shown in subgroup analyses of wait list controls (SMD -0.63, 95% CI -0.83 to -0.45; nine comparisons) and controls that received treatment as usual (SMD -0.23, 95% CI -0.37 to -0.09; seven comparisons). The results were not substantially changed in sensitivity analyses that only included only comparisons with neither the Beck Depression Inventory I or II as the primary outcome measure (SMD -0.32, 95% CI -0.48 to -0.17; seven comparisons), comparisons where acceptable attrition rates (<20%) were reported (SMD -0.59, 95% CI -0.85 to -0.34; seven comparisons), comparisons without significant differences in attrition rates between CCBT and control groups at intervention endpoint (SMD -0.50, 95% CI -0.73 to -0.27; nine comparisons) or comparisons reporting modern imputation techniques (SMD -0.34, 95% CI -0.51 to -0.18; eight comparisons).

At long-term follow-up (>6 months), no statistically significant difference in depression symptoms were found between the groups (SMD -0.05, 95% CI -0.19 to 0.09; five comparisons; Ι²=7%).

At intervention endpoint, no statistically significant differences between groups were found in relation to function improvement (SMD -0.05, 95% CI -0.31 to 0.22; 12 comparisons; Ι²=85%). Rates of attrition were statistically significantly higher among CCBT patients compared with controls (RR 1.68, 95% CI 1.31 to 2.16; 16 comparisons; Ι²=55%). Evidence of publication bias was found (Begg's test p=0.009; Egger's test p=0.01; trim-and-fill method SMD -0.32, 95% CI -0.49 to -0.16).

Authors' conclusions

Despite a reduction in depression at post-treatment, the effect at long-term follow-up and the function improvement were not significant, with substantially high dropout. The clinical usefulness of current CCBT for adult depression may need to be re-considered downwards in terms of practical implementation and methodological validity.

CRD commentary

The review question and inclusion criteria were clearly defined. Relevant databases were searched and attempts were made to locate unpublished literature. The search was performed up to 2011 and it was possible that additional relevant studies were published after that. It was unclear whether any of the review processes were performed in duplicate, so reviewer error and/or bias could not be ruled out. Some quality assessment criteria were stated as part of the review's inclusion criteria; these did not cover all of the quality-related issues that can be present in RCTs, so there was potential for within-study biases.

Study characteristics were presented in sufficient detail but it was unclear whether double counting of control group data may have occurred for the two studies that reported multiple comparisons. Attempts were made explore the impact of study differences (such as control type, attrition rate, main rating scale used) via subgroup and sensitivity analyses. The authors acknowledged that they were not able to include data from unpublished studies (reasons not provided).

The authors' conclusions reflect the evidence presented and appear reliable; more recent studies may have become available since the review search was performed in 2011.

Implications of the review for practice and research

Practice: The authors stated that the risk of overestimation when using self-rating scales (including the Beck Depression Inventory) should be borne in mind when CCBT is adopted for clinical use. The clinical usefulness of current CCBT for adult depression may need to be re-considered downwards in terms of practical implementation and methodological validity.

Research: The authors stated that research designs with wait list controls are prone to overestimation of the intervention; this issue should be more carefully considered in general RCT settings, as well as in RCTs of CCBT. Research on CCBT should give more consideration to the influence of imputation methods on results.


No external funding.

Bibliographic details

So M, Yamaguchi S, Hashimoto S, Sado M, Furukawa TA, McCrone P. Is computerised CBT really helpful for adult depression? A meta-analytic re-evaluation of CCBT for adult depression in terms of clinical implementation and methodological validity BMC Psychiatry 2013; 13(1): 113. [PMC free article: PMC3638010] [PubMed: 23587347]

Indexing Status

Subject indexing assigned by NLM


Adult; Cognitive Therapy /methods; Depressive Disorder /psychology /therapy; Humans; Therapy, Computer-Assisted /methods; Treatment Outcome



Database entry date


Record Status

This is a systematic review that meets the criteria for inclusion on DARE.

CRD has determined that this article meets the DARE scientific quality criteria for a systematic review.

Copyright © 2014 University of York.

PMID: 23587347

PubReader format: click here to try

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...