This review concluded that peer support interventions increased breastfeeding continuation in low- and middle-income countries, but not in high-income countries (UK included). Low-intensity interventions did not seem to be effective. These conclusions reflect the evidence presented and seem appropriate.

To examine the effect of setting, intensity and timing on the effectiveness of peer support interventions to encourage breastfeeding.

The authors searched BNI, CINAHL, The Cochrane Library, EMBASE, MEDLINE and an unspecified trials website to June 2011. Search terms were reported online. There were no language restrictions. Reference lists of retrieved articles were searched manually.

Randomised controlled trials (RCTs) that recruited pregnant or postpartum women and compared peer support with usual care were eligible for the review. Peer support was defined as support offered by women who had received suitable training and either had experience of breastfeeding or came from a similar background to the women they were supporting. Support could be provided in the antenatal and postnatal periods or in the postnatal period only. Outcomes of interest were any breastfeeding and exclusive breastfeeding at the end of follow-up. Follow-up needed to be at least four weeks.

Settings of the included trials covered a range of low-, middle- and high-income countries. The number of planned contacts to provide support ranged from one to 10 or more. Interventions covered the antenatal and postnatal periods in about half of the trials.

Two reviewers independently assessed studies for inclusion; disagreements were resolved in consultation with other reviewers.

Two reviewers independently assessed quality using the methods of the Cochrane Handbook for randomisation, allocation concealment, description of withdrawals and drop-outs, blinded outcome assessment and selective outcome reporting.

Relative risks (RRs) of not breastfeeding or not exclusively breastfeeding at the end of follow-up were extracted or calculated, together with their 95% confidence intervals (CIs). It was assumed that participants lost to follow-up had stopped breastfeeding.

Two reviewers independently extracted data; disagreements were resolved in consultation with other reviewers.

Studies were pooled by meta-analysis using random-effects models. The design effect was calculated for cluster RCTs and used to make appropriate allowance for clustering. Statistical heterogeneity was assessed using the Χ² test and I² statistic. Prespecified subgroup and meta-regression analyses were used to explore the effect of setting (high- versus middle- or low-income countries), intervention intensity (fewer than five versus five or more contacts) and timing (antenatal and postnatal versus postnatal only) on the effectiveness of the intervention. RCTs performed in the UK were analysed separately.

Seventeen trials were included in the review and 15 of these contributed data to the quantitative synthesis. Several trials did not give enough information to assess risk of bias fully. Last study follow-up was at three to six months for all except two trials. Losses to follow-up ranged from 1% to 41% but were generally balanced between groups. Only five trials reported both the number of contacts received and the proportion of women in the intervention group who received some support.

Overall, peer support reduced the risk of not breastfeeding by 15% (RR 0.85, 95% CI 0.77 to 0.94) and of not exclusively breastfeeding by 18% (RR 0.82, 95% CI 0.76 to 0.88). Both effects were statistically significant. Significant heterogeneity was present for both outcomes. In UK RCTs, peer support did not significantly reduce the risk of not breastfeeding (RR 0.96, 95% CI 0.89 to 1.04; three RCTs) and not exclusively breastfeeding (RR 0.98, 95% CI 0.96 to 1.01; two RCTs).

Subgroup and meta-regression analyses showed that peer support had a significantly greater effect in low- or middle-income countries than in high-income countries for both outcomes. Giving support at higher intensity and in the postnatal period had a statistically significant effect for any breastfeeding and not for exclusive breastfeeding.

Peer support interventions increased breastfeeding continuation in low- and middle-income countries, but not in high-income countries (particularly the UK). Low intensity interventions did not seem to be effective.

The review question and inclusion criteria were clear. The search covered various relevant sources and there were no language restrictions. Some efforts were made to locate unpublished trials, but publication bias was not assessed. Study selection, quality assessment and data extraction were done by two reviewers independently, which minimised risks of errors and bias affecting the results.

Appropriate methods were used to assess study quality and the results were used in the synthesis. Prespecified analyses were used to investigate differences between trials. A limitation of the review was that poor reporting in the included trials meant that the analysis was based on the interventions as planned rather than as actually delivered.

The authors' conclusions reflect the evidence presented and seem appropriate, although it should be borne in mind that subgroup analyses based on summary data can sometimes produce spurious results. Effectiveness of peer support in UK settings was specifically investigated and again the authors' conclusions and recommendations seem appropriate.

Practice: The authors stated that peer support interventions should be recommended for low- and middle-income countries, but may not be effective where routine services to support breastfeeding are already established.

Research: The authors stated that any new peer support services in high-income countries should only be implemented with a robust evaluation of outcomes.

Part funded by the National Institute for Health Research CLAHRC (Collaboration for Leadership in Applied Health Research and Care) for Birmingham and the Black Country.

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.