The review found that peer support interventions for depression resulted in greater improvement in depression symptoms than usual care and may have similar efficacy to group cognitive-behavioural therapy. The authors' conclusions reflected the evidence base, but heterogeneity in the participants and interventions means that the reliability of the conclusions is not clear.

To assess the efficacy of peer support compared to either usual care or psychotherapy for reducing symptoms of depression.

MEDLINE, PsycINFO, CINAHL and Cochrane Central Register of Controlled Trials (CENTRAL) were searched up to April 2010 to identify relevant published studies; search terms were reported. Reference lists of retrieved studies were searched.

Eligible studies were peer reviewed randomised controlled trials (RCTs) where adult participants with current symptoms of depression were assigned to either a peer support intervention or a usual care or psychotherapy control condition. Symptoms were required to be measured before and after the intervention in both peer support and control groups.

In the included studies, the mean age of participants ranged from 20 to 74 years and the proportion of female participants ranged from zero to 100%, where reported. Participants were highly heterogeneous and included women in the perimenopause, with young children at home or postpartum, males with HIV, either gender with cancer, caregivers of sick patients and elderly patients after admission for psychiatric care. Control groups received either usual care or group cognitive-behavioural therapy (CBT). Depressive symptoms were measured by the Beck Depression Inventory (BDI), Centre for Epidemiological Studies Depression Scale, Zung Self-Rating Depression Scale, Hamilton Rating Scale for Depression (HRSD) and the Edinburgh Postnatal Depression Scale. Peer support group interventions ranged from once a week for nine months to 12 sessions weekly; one intervention used a telephone peer mentor for 10 weeks.

Two reviewers selected studies for the review. Discrepancies were resolved by discussion and consensus.

Studies were assessed for quality using criteria similar to the Jadad scale; criteria included presence and quality of randomisation, blinding of raters, adequate accounting and percentage of drop-outs.

The authors did not state how many reviewers assessed the studies for quality.

Data were extracted on mean differences between baseline and post-intervention scores on the depression scales, together with 95% confidence intervals (CIs), onto a standardised data extraction form. The mean differences were used to calculate standardised mean differences (SMDs).

The data extraction form was developed and tested by one reviewer and reviewed by a second reviewer. The authors did not state how many reviewers performed data extraction.

A DerSimonian-Laird random-effects model was used to pool standardised mean differences and calculate 95% CIs. Separate analyses compared peer support to usual care and compared peer support to CBT. Stratified analyses were undertaken according to gender, type of depression scale used and blinding of raters to measure outcomes. Heterogeneity was assessed with the X² test and quantified by the I² value. Between-study variance (Τ²) was calculated. Influence analyses were conducted to assess the effects on the results of exclusion of individual studies one at a time. Intention-to-treat analyses of all randomised participants were conducted by assuming baseline scores of drop-outs remained constant. Publication bias was assessed by inspection of contour-enhanced funnel plots and using the Begg and Mazumdar adjusted rank correlation test and the Egger regression asymmetry test.

Ten RCTs (1,269 participants) were included in the review. Three trials compared peer support to usual care. Two trials compared peer support to CBT and five trials compared peer support to both usual care and CBT; in one trial participants were not randomised to usual care and so findings from this group were not included in the review. All studies were randomised and described participant drop-outs. None of the studies wwere double blinded, three studies blinded outcome assessment and three studies described the method of randomisation. Study completion rates ranged from 59% to 98%.

Peer support (seven studies) was associated with a statistically significant reduction in depressive symptoms when compared to usual care (SMD -0.59, 95% CI -0.98 to -0.21; seven studies, 849 participants) with substantial heterogeneity (I²=71%). Stratification by gender and depression measure did not markedly alter heterogeneity, but heterogeneity was reduced when subgroups were stratified by blinding of raters. Intention-to-treat analysis confirmed the findings of the overall analysis and the findings were not changed by influence analysis. There was no evidence of publication bias.

There was no evidence of a significant difference in depression outcomes between peer support and CBT (SMD 0.10, 95% CI -0.20 to 0.39; seven studies, 301 participants) with moderate heterogeneity (I²=40%). Results were similar in subgroups with blinded and unblinded raters. The findings were not markedly changed with influence analysis. There was no evidence of publication bias.

Peer support interventions for depression resulted in greater improvement in depression symptoms than usual care and may have similar efficacy to group CBT.

The review addressed a clear research question. The broad inclusion criteria were appropriate. A range of relevant sources were searched for peer reviewed published studies with no details of language restrictions; language and publication biases could not be ruled out. Appropriate methods were used to select studies for the review. It was not clear how many reviewers assessed studies for quality or extracted data, so reviewer error and bias could not be ruled out. The eligible studies included a wide range of participants and interventions, which made it difficult to generalise the findings. Outcomes were measured appropriately by validated depression scales, but only three studies blinded their assessors. The authors appropriately used standardised mean differences to adequately measure symptoms assessed with different depression scales. Synthesis of studies in meta-analyses, stratification to explain potential statistical heterogeneity, sensitivity analyses to investigate the influence of individual studies and intention-to-treat analyses to confirm the stability of the results were all appropriate. Blinded studies yielded smaller positive effects than unblinded studies, which the authors suggested may indicate bias among these lower-quality studies. There was no evidence of publication bias by formal tests, but very few studies contributed to this analysis. The authors acknowledged that the evidence was too limited to assess the effectiveness of different types of peer support, frequency, duration and training requirements.

The authors' conclusions reflected the evidence base, but heterogeneity between participants and between interventions means that the reliability of the conclusions is not clear.

Practice: The authors stated that peer support may be a cost effective alternative when other depression care interventions are unavailable, unaffordable and unacceptable.

Research: The authors stated that larger RCTs of patients with acute and chronic depression in primary care settings were needed. Research should focus on determining the most effective methods, duration and frequency of peer support and training, together with costs and mechanisms for providing benefits.

VA Health Services Research and Development Service; Michigan Diabetes Research and Training Centre; Michigan Institute for Clinical and Health Research.

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.