A meta-analysis of education based breastfeeding interventions: Impact of social marketing techniques, number of intervention components used, and methodological quality

The literature reflects considerable variation in the extent to which educational breastfeeding interventions are found to be effective. Three factors were hypothesized to account for this variation: the use of social marketing (which was defined as including audience segmentation and audience/intervention research), the number of intervention components used, and the methodological quality of the individual studies. A random effects meta-analysis was used to examine the impact of these variables using the data from 52 studies. Separate analyses were performed for each of three temporal outcome categories: initiation, short-term duration, and long-term duration. Type of breastfeeding measured and type of control group used were controlled for as variables.; Educational breastfeeding interventions had a significantly positive impact on effect size magnitude in all analyses, across all three temporal outcome categories. A sensitivity analysis was performed using data from only randomized controlled trials (RCTs). Although still significant, RCT effect sizes were smaller than the effect sizes from the entire pool of studies. A second sensitivity analysis was performed on only data from studies that used an alternative treatment control group. Although mostly still significant, effect sizes from studies that used an alternative treatment control group were smaller relative to effect sizes from the whole group of studies. The one case in which the alternative treatment control group effect size was not significant was long-term duration outcomes in RCTs. Educational interventions may not be more effective than standard treatment alone for long-term duration outcomes.; RCT analyses indicated that the use of social marketing improves both initiation and long-term duration success rates. At p=.07, social marketing narrowly missing having a significant impact on short-term duration outcomes. The use of multiple intervention components (especially support) improves long-term duration outcomes. Some evidence indicates that social marketing and multidimensionality are correlated. Worse quality studies are prone to finding spuriously inflated short- and long-term duration outcomes.; It is suggested that future analyses further examine the relationships among predictor variables. Future research should also control for other relevant factors and include more thorough and standardized reporting. Literature interpretation and reviews should rely on high quality studies whenever possible.