Dear Beth:

First, JAMA is not know for thorough peer review.  It is the type of journal that you choose if you want your research to get out there fast.  

Second, if this is part of the larger studies of the "intervention" in Belarus -- it is quite problematic in that the "intervention" was only partially successful.  This means that not everyone complied with the intervention and some in the nonintervention group might have improved as well.  It is not as clean as giving a drug with a placebo -- you have a lot of overlap.  This means that you completely water down the ability to detect a difference.  All this study showed is that the INTERVENTION to improve breastfeeding rates was not successful enough to show statistically significant differences.  It did NOT test the efficacy of breastfeeding itself.  So, say for instance you did a randomized trial that showed a drop of 30% in childhood mortality in areas of vitamin A deficiency in a randomized trial with placebos and vitamin A capsules (these are real stats by the way).  Then you did an intervention trial where you distributed those capsules to some communities and not others.  But some of the people in the communities that didn't get the vitamin A capsules snuck over to the communities where they were being given and some of the people in the communities that got vitamin A capsules didn't take them because they thought they would cause babies to die (also a real example).  Then in this instance, you would have a drop in mortality rates in the communities where people were not supposed to get the vitamin A capsules and you would also see less of a drop in mortality rates in the communities where people were supposed to get the vitamin A capsules but didn't.  So, you might see a drop of 10% -- BUT because the intervention was at a COMMUNITY level and NOT the INDIVIDUAL level -- your sample size is based on the community -- not the individual and voila -- your statistical power is lower and the result is no longer significant.  

 I don't have access to the full article and its too expensive for me to download.  Nevertheless - if you look at the stats -- they are only talking about breastfeeding to 3 months.  Hardly what I would call "longer term breastfeeding".  Also -- unless they really looked into the phenomenon of supersizing bottles of breast milk -- I would argue that you aren't really testing one of the key factor contributing to obesity.  Then because obesity is multifactoral -- if the family diet were very good you might see the impact of overfeeding disappear initially only to reappear in old age. 

Finally, you can NEVER claim that there was proof of NO effect --- you can ONLY prove that there is a statistical PROBABILITY of that an intervention actually did result in the expected result.  


Best, Susan Burger

             ***********************************************

Archives: http://community.lsoft.com/archives/LACTNET.html
To reach list owners: [log in to unmask]
Mail all list management commands to: [log in to unmask]
COMMANDS:
1. To temporarily stop your subscription write in the body of an email: set lactnet nomail
2. To start it again: set lactnet mail
3. To unsubscribe: unsubscribe lactnet
4. To get a comprehensive list of rules and directions: get lactnet welcome