More information about text formats
Congratulations to Bland et al. for a well-done study, well analyzed and well written up on an issue that needed to be studied the way they did it. I agree with most of their conclusions, particularly their warning that studies linking postnatal HIV transmission to feeding practices need to use more exacting methods to document the latter than most studies have used to date.
I would add to...
I would add to this another point: the "standardised WHO definitions" that they refer to were developed as a way of ensuring that household survey results are comparable and are not necessarily appropriate for studying the relationships between feeding patterns and certain types of morbidity. Thus the fact that the WHO definition "allows" "drops or syrups consisting of vitamins, mineral supplements, or prescribed medicines" should not be accepted unthinkingly. All of these can be provided in a contaminated or outdated form and all contain additional substances that are used as preservatives, fillers, binders, etc. While this may not matter much when studying for example the relationship between feeding patterns and concurrent patterns of diarrhea or growth, it may make a big difference when relating feeding patterns to such outcomes as allergy, micronutrient absorption--and HIV transmission.
The population Bland et al. studied is likely fairly representative of feeding patterns in wealthier rural African settings. However, in other settings their findings are less likely to hold. More importantly, their population was probably not the right one in which to compare the various measurement methods they used. The main reason for this is that there was too little variability in the duration of exclusive breastfeeding in their population. Only 17% were still exclusively breastfeeding after two weeks of age and thus able to switch category. That is, after this age so few infants change categories that high levels of sensitivity and specificity in the tools used to measure it were in a sense unavoidable. This is not to fault their study, but to point out that it needs to be repeated in other settings before generalizing from their findings.
Finally, that mothers tended to exaggerate the duration of EBF when recalling it several months later could in part be an artefact of the slightly different method of operationalizing the definition of EBF that was used in the long-term recall. Water was the only non-milk fluid used and the extent to which other fluids (or fluids mothers did not define as "water" such as “glucose”) were given before water could be partially explanatory.
(1) RM Bland, NC Rollins, G Solarsh, J Van den Broeck, and HM Coovadia. Maternal recall of exclusive breast feeding duration. Arch Dis Child 2003; 88: 778-783 .