Ten RCTs (1,445 participants) were included. Sample sizes ranged from 48 to 583. All included RCTs used a prospective design, gave an adequate description of study outcome, used adequate statistical methods, reported results relating to the primary outcome, and interpreted results which were deemed to be generalisable. Nine RCTs reported appropriate outcomes and confounding factors. Only three RCTs reported using an intention-to-treat analysis.
Educational programmes were effective for promoting the initiation of any form of breast feeding in low-income mothers (RR 1.46, 95% CI 1.03 to 2.08; seven RCTs; Ι²=86%), and reported greater effects for exclusive breast feeding (RR 1.72, 95% CI 1.34 to 2.21; four RCTs; Ι²=50.5%). Programmes that encouraged mothers to continue breast feeding also reported significant rates at three to six months postpartum (RR 1.15, 95% CI 1.01 to 1.30; five RCTs; Ι²=0%). But there were no significant differences between intervention and control groups for breast feeding at six weeks to two months post partum (seven RCTs). The successful programmes usually involved multiple short (less than 20 minutes to 30 minutes) follow-up appointments (four RCTs).