Twenty-five RCTs were included in the review (n=8,942 participants, sample size ranged from 56 to 2,757).
Breastfeeding (six RCTs): Three trials reported statistically significant positive results for increased breastfeeding amongst new mothers, but the overall pooled result showed no statistically significant benefit. There was considerable heterogeneity (p<0.001) between the trials.
Physical activity (five RCTs): Peer-based interventions lead to a statistically significant positive effect (SMD 0.16, 95% CI 0.05 to 0.27; three RCTs). There was no evidence of heterogeneity (p=0.57) between the trials.
Medication adherence (three RCTs): Only one trial reported a statistically significant result. The pooled result from the two trials reporting odds ratios found no overall statistically significant benefit. There was significant heterogeneity between trials (p=0.049).
Women's health cancer screening (three RCTs): All three trials reported increases in either mammography or gynaecological cancer screening, but the pooled result was not statistically significant. There was considerable heterogeneity (p<0.001) between the trials.
Smoking cessation (two RCTs): Both trials reported higher levels of smoking cessation for the intervention group, and the pooled result showed a statistically significant benefit for the intervention group (OR 1.64, 95% CI 1.09 to 2.46), with no evidence of heterogeneity (p=0.47).
Other outcomes: Two trials assessed an increase in general activities in participants with chronic disease, but overall there was no statistically significant difference between the peer-based intervention and the control. Two trials assessed condom use; the pooled result showed statistically significant increases for the intervention group (OR 2.27, 95% CI 1.15 to 3.54), with no evidence of heterogeneity (p=0.80). One trial assessed advanced directives among dialysis patients; it reported a statistically significant increase for the intervention group. One trial reported moderate but statistically significant changes for the intervention group for a reduction in alcohol intake among university students.
Subgroup analyses resulted in similar effect sizes, but did not appear to explain any of the observed heterogeneity.