Weighing the Evidence: Wherefore Meta-Analysis and Odds Ratio?

Abstract

Are there any pediatricians who doubt that “breast is best?” When counseling our families about breastfeeding benefits, however, we should try to quantify the benefits. A meta-analysis summarized in this issue offers evidence that lack of breastfeeding in infancy is associated with a higher risk of leukemia in children. However, the limitations of meta-analyses and odds ratios, in particular, bear noting. Meta-analysis is a powerful statistical method by which data from different studies can be combined to make a stronger conclusion than could be made from the individual studies themselves. If misapplied, it can result in powerfully wrong answers, but the authors of this article conducted a careful search for publication bias and heterogeneity of studies, lending credence to their results. What about the mother who cannot breastfeed? …