6Future Research

Publication Details

  • It is conceivable that a cluster randomized study similar to the PROBIT study in Belarus could still be undertaken in this country, as Baby Friendly Hospital Initiative (BFHI) is not yet widely adopted (only 1.3% of the maternity units in this country is designated Baby Friendly (http://babyfriendly.org/, accessed 6-7-2007)). Such a study is important because the magnitude of effects measured from such an intervention is useful in assessing the public health impact in a socio-cultural environment that is not as breastfeeding friendly as the one in Belarus. It should also be noted that studies in the literature reported good success in improving the initiation rate of exclusive breastfeeding in hospitals that had achieved the Baby Friendly status, but those rates declined rapidly after discharge. Thus, if such a study is undertaken, step number ten of the BFHI ten steps, postdischarge breastfeeding support, must be designed carefully and implemented.
  • For future studies on the effects of breastfeeding interventions, it may be preferable to focus on the rate of exclusive breastfeeding rather than the rate of any breastfeeding. The larger effects seen with exclusive breastfeeding compared with any breastfeeding in our meta-analyses suggested that the widely varying classifications of breastfeeding exposures in the any breastfeeding category might have biased the findings toward the null effect.
  • Our results suggest that prenatal combined with postnatal interventions could be effective in prolonging the duration of breastfeeding. Future studies on particular interventions should take this possibility into account and emphasize interventions in both the prenatal and postnatal periods.
  • In our overall analysis, we did not find that professional support was effective in increasing the rate of breastfeeding initiation or duration but we found that lay support was effective in increasing the rate of short- and long-term breastfeeding. It may be instructive to compare the two forms of support in a head-to-head trial to further understand the similarities and differences so that better breastfeeding support could be designed and implemented.
  • One fair quality study on postpartum skin-to-skin intervention reported a 2 months increase in breastfeeding duration compared to usual care but the number of participants who received the intervention in the study was small (N=30). It would be desirable to confirm the effect of postpartum skin-to-skin intervention on breastfeeding duration in a larger trial.
  • One fair quality study on prenatal breastfeeding education for expectant fathers reported a significant increased rate of full breastfeeding at 6 months compared to infants whose fathers did not receive such training. It would be important to conduct a head-to-head trial where the fathers were directly randomized to intervention or control. This will lend confidence to the effects reported. More studies involving other family members (e.g., grandparents, partners) will be of value to clarify the effects of interventions to promote and support breastfeeding.


The final report was submitted on 7-27-2007 to AHRQ. On 8-20-2007, we were alerted by David Meyers, M.D. of a 2005 study concerning fathers and breastfeeding success63 that was not included in our report. We reviewed the study and found that it met our inclusion criteria. The final report was therefore revised to include this study but the overall meta-analysis was not re-conducted. An examination of our original literature search strategy including the terms “breastfeeding” and “controlled trials” did not reveal an apparent reason for the inadvertent omission of this study.