The review concluded that although more studies were needed, internet and computerised treatments had promise as potential treatments for depression. The review had some methodological problems, but the authors’ conclusions were suitably cautious and appear appropriate.

To summarise the effects of internet-based treatments for depression and investigate the characteristics of studies that may be related to the effects.

EMBASE, PubMed, PsycINFO and Cochrane Central Register of Controlled Trials (CENTRAL) were searched from 1966 to January 2009 for articles in any language. Reference lists of included studies relevant reviews were searched.

Randomised controlled trials (RCTs) of internet-based or computerised psychological treatments versus noncomputerised psychological interventions or control in adults with depression (with or without anxiety) were eligible for inclusion. Trials of patients with comorbid medical or psychological conditions were included. Depression could be established with a diagnostic interview or elevated levels of depressive symptoms based on self-reported measures. Trials on in-patients or children (<18 years) were excluded. Component studies were included.

The included trials compared internet-based/computerised psychological interventions (predominantly internet-based cognitive-behavioural therapy) with control or face-to-face psychological interventions in adults with depression. Depression prior to trial entry was established in a variety of ways, although most studies did not include a diagnosis based on a formal diagnostic interview. Most trials were conducted in adults in the community. Trial durations, where reported, ranged from five to 11 weeks. Most trials provided additional therapist support, including help at the beginning and/or end of the trial and additional contact via face-to-face, email, telephone or postcard. The reported outcomes included Beck Depression Inventory (BDI), Centre for Epidemiological Studies Depression scale (CES-D), Hamilton Depression Rating scale (HAM-D) and Major Depression Inventory (MDI).

The authors did not state how many reviewers performed study selection.

Quality assessment was undertaken using the Cochrane four-point scale of allocation concealment, blinding, completeness of outcome data and randomisation. Results from the validity assessment were not provided.

The authors did not state how many reviewers performed the validity assessment.

Data on depression and anxiety measures were extracted and used to calculate post-test effect sizes (ES) and 95% confidence intervals (CIs). The mean of the effect size was calculated for studies that used more than one measure of depression, so that each study provided only one effect size.

Pooled effect sizes were calculated using a random-effect meta-analysis. Statistical heterogeneity was assessed using I² and the Cochrane Q test. Subgroup analysis was undertaken according to type of intervention and comparator, psychological treatment, anxiety allowed and professional support. Publication bias was assessed using funnel plot analysis and the effects were explored using Duval and Tweedie's trim-and-fill method. Sensitivity analysis omitted outliers and removed multiple comparisons from the study.

Twelve RCTs were included in the review (n=2,446). Sample sizes ranged from 30 to 343 participants. The authors reported that the trial quality was not satisfactory.

Compared with control, internet-based psychological interventions were statistically significantly superior (ES 0.41, 95% CI 0.29 to 0.54, I²=57%). Subgroup analysis indicated that professional support had a significant impact upon the results. Studies that had professional support alongside the intervention had a higher effect size (ES 0.61, 95% CI 0.45 to 0.77, I²=24%) compared with those that offered no professional support (ES 0.25, 95% CI 0.14 to 0.35, I²=10%).

There was no evidence of publication bias.

Although more studies were needed, internet and computerised treatments had promise as potential treatments for depression.

Inclusion criteria for the review were clearly defined. Several relevant databases were searched for articles in any language. Publication bias was assessed and not detected. The authors did not state whether they made attempts to minimise error and bias throughout the review process. Quality assessment was based on a simple checklist of trial biases and was not reported, although the authors acknowledged the unsatisfactory quality of the data. Trials were pooled using a random-effects meta-analysis and subgroup analysis was undertaken, which was appropriate.

Overall, the review had some methodological problems, but the authors’ conclusions were suitably cautious and appear appropriate.

Practice: The authors did not state any implications for practice.

Research: The authors stated that there was a need for studies with longer follow-ups and studies conducted in patients in psychiatric settings. Future trials should have accurate diagnostic procedures and explore effects in different patient populations.

Swedish Research Council.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.