there seem to be a lot of people unhappy with the study on breastfeeding and
asthma. I have not read the study and just read the abstract quickly.
HOWEVER!
This asthma study is a spin off from PROBIT, which was warmly welcomed here
on lactnet as a rather ground-breaking study.
One of the problems with research into the outcomes of different types of
infant feeding is the puzzle of how to acheive robust quality studies. RCTs
are typically considered to be gold standard research, with observational
studies only ever less robust.
Yet, would it be ethical or practical to conduct an RCT of infant feeding???
For a long time those in the field held that it was not a good idea. In
2000 the only RCT I know of which randomised infant feeding -- a study by
Nduati et al on HIV transmission -- was published. In this study we saw
problems with the uptake by eligble women (18% asked agree to take part) and
therefore problems with whether this would apply to any real-life
population. (There were other problems discussed at the time).
Kramer et al, in the PROBIT study sought to approach the problem from a
novel, and rather elegant, way, by randomising at the level of
hospital/clinic and delivering an intervention of baby friendly care versus
normal care. Belarus was chosen in part because the heirarchical structure
of the health services there meant that the country could commit to having
randomly chosen hospitals go BFI within a year, on the say so of the head of
the facility (not a condition that would be met in many countries /
settings).
At the same time as collecting data on the effect on the two populations --
women delivering at BFI facilities and areas implementing 'usual care',
PROBIT researchers collected data on growth and on asthma incidence.
It may be (I stress, I have not yet read the paper) that this was not a
great design to study asthma, or that the way the conclusions are worded is
a bit unhelpful. However, the design of an rct requires that data is
analysed between the groups on the basis of randomisation. To analyse sub
sets -- eg of babies who were ebf for so long -- is not as robust a way to
handle the data. So the researchers chose their data analysis method before
analysing -- this is PROPER!!
Now, the original PROBIT study is the real workhorse study which shows the
difference BFI makes, and I know I have quoted it over the years in many
contexts. It is one of the benchmark pieces of research that shows that
implementing BFI is worth getting out of bed for, and it does it in an RCT.
I have also found the growth data from PROBIT of great interest, including
the speculations about the relationships between babies' gowth and mothers'
feeding choices. For me, if this asthma study is not so great, (and I am
not saying that, I don't know) , I am willing to cut the research team a lot
of slack for the other great stuff that has come out of their data.
I am not saying that the study should not be critiqued, but I am beginning
to feel as if this study is being a bit demonised (perhaps on the basis of
the inept press reports? Perhaps because it says something we don't like to
hear???). There is another big, UK, feeding study which generated a lot of
different aspects to the data and I wish some of the questions had been
better framed, etc etc. But I am grateful that the researchers tried to
gather data on several outcomes, since they were doing a large, expensive
piece of research -- something which is hard to set up and fund. Maybe
there are aspects of this study that will be useful to other asthma/infant
feeding researchers. Rare indeed is the study which has the 'answers', its
a piecemeal job.
Having done research myself, I think the greatest wish I had was that people
would believe that I conducted it honestly and in the spirit of open enquiry
of my data. I would like to extend that to these researchers, because I
have so valued their other work on PROBIT, on growth, and Kramer's part in
the review of optimal duration of breastfeeding which informed the WHO in
changing the recommendation from 4 to 6 months. I know this sounds
Pollyannaish -- not something I am usually accused of!!!-- am I the only one
who is thinking this way???
thanks
Magda Sachs, PhD
***********************************************
Archives: http://community.lsoft.com/archives/LACTNET.html
Mail all commands to [log in to unmask]
To temporarily stop your subscription: set lactnet nomail
To start it again: set lactnet mail (or [log in to unmask])
To unsubscribe: unsubscribe lactnet or ([log in to unmask])
To reach list owners: [log in to unmask]
|