Skip to content

Is breast truly best?: the media use a single study to reject decades of research

by on 2014/03/05

There has been a flurry of media activity in the last week about the merits of breast versus formula milk following the publication and associated press release of a U.S. study that examined whether breastfeeding is associated with better health and academic outcomes. It is fair to say that the media in general have gone to town in their interpretations of this study with headlines such as breast milk is ‘no better for a baby than bottled milk and it INCREASES the risk of asthma’ and ‘Hold the guilt! New study finds benefits of breastfeeding dramatically overstated’. Such headlines question decades of research which has shown the benefits of breast milk for infants (see WHO and NHS guidelines) so what did this study do that enabled the media to make such extreme statements?

What did the study involve?

The study was conducted by Cynthia Colen and David Ramey from the Department of Sociology at Ohio State University and it was published in the Journal Social Science and Medicine. The study has been comprehensively described by the NHS. This was a cohort study and data was used from the U.S. National Longitudinal Survey of Youth (NLSY). The researchers compared aspects of physical health (body mass index, obesity and asthma), behaviour (hyperactivity, attachment and behavioural compliance), and academic achievement (reading comprehension, memory, vocabulary and academic performance) between the ages of 4 and 14 years in 8,237 children. In the first stage of the analyses they compared breastfed and formula fed children on these measures. In the second stage, they compared siblings (1,773) where one was breastfed and another was formula fed. Using this design they were able to control for socio-economic status, maternal education and employment and other variables that may act as confounding factors (factors that may influence the study findings that are not directly examined in the study) in these types of studies.

What did the study find?

While breastfed children had better outcomes on a wide range of measures when data from the entire sample was analysed, these differences disappeared when siblings were compared where one was breastfed and one was formula fed.   The researchers concluded that the “relationship between breastfeeding and long-term childhood outcomes may not be as consistent and straightforward as once thought”. The researchers infer that in other studies socio-economic status and other parental factors may have driven findings of better outcomes in breastfed children rather than the breast milk itself.

Are there any methodological issues that limit interpretation of the findings?

First, the measures of breastfeeding used in the study can be considered crude. The researchers simply divided the sample into breast and formula fed on a categorical basis of whether they were ever breastfed or not. This means that a child who was only breastfed for one day will be in the same category as a child exclusively breastfed for a year. Despite the known importance of colostrum (the initial substance the baby receives when breastfeeding, before the mother’s milk comes in) no health professional would argue that the health benefits of a day or so of breastfeeding would impact significantly on a child such that they would differ from their sibling who did not receive this.   The outcome measures of this study, e.g. hyperactivity and academic achievement are in no doubt affected by multiple factors.

Second, the critical findings – where no differences were reported between formula and breastfed children – were based on a sibling comparison. The authors neither report nor control for the reasons why a mother breastfeeds one child and not another. For some of these mothers it may have been that they encountered problems with breastfeeding and gave up early on moving to formula feeding directly for a second child. Without this information it is impossible to tell if such biases existed in the sample. These kind of biases would mean that the sibling comparison was comparing children who over the course of infancy were in fact fed very similarly, which makes it difficult to understand why you would expect to see differences in health and other outcomes over a 10 year period?

Third, there was no measurement of exclusivity included – many children are combined fed (breast and formula) and such children were presumably included in the breastfed group. There is no reporting or controlling for when solid food is introduced either – the timing of which varies from individuals to individuals and will impact on quantity of breastfeeding.

Fourth, while the researchers included a number of outcomes across areas of physical health, behaviour and academic achievement, they did not include a range of other outcome measures for which the strongest and most consistent effects of breastfeeding have been shown, such as allergies and diabetes.

Are the findings relevant to mothers outwith the U.S.?

In the U.S., unlike the U.K. and many other European countries, there is limited availability of paid maternity leave which will impact on breastfeeding choices. The findings should be treated with caution in terms of generalising to a non-U.S. population where maternity leave is available to all.

Did the media misrepresent the findings?  

The Daily Mail tried to have it’s cake and eat it – their headline was that breastfeeding increased the risk of asthma but this (and it was an association not a causal link) was only reported in the whole sample. So, while this newspaper felt it was able to make extreme statements about breastfeeding based on the importance of the sibling data and dismissing the significant positive effects of breastfeeding in the whole sample data, they did not extend this dismissal to the negative asthma finding from the whole sample data. It needs to be remembered that this is a cohort study and can only show associative links between variables. It has been suggested that it may be that in cases where asthma runs in the family the mother is more likely to breastfeed to attempt to reduce the risk of her infant developing asthma and therefore more end up in this category (see this response from a paediatrician to the asthma findings). The Daily Mail also repeatedly used the term ‘expert’ to refer to the researchers. A quick glance at their credentials (see above links) reveals that they are not experts in breastfeeding as suggested by the news article. The use of this term was no doubt to heighten the impact of the conclusions made. Only last week we at Research the Headlines wrote about an other article in the Daily Mail in which a Professor of Evolutionary Biology, Katie Hinde, was completely misquoted about comments she made following a conference presentation about the quality of breast milk. Together the articles suggest the tendency for a bias in reporting on the issue of breastfeeding.

It is highly concerning that the media have by and large written off the positive impact of breastfeeding, for infant and mother, on the basis of one study with potential confounds that may have contributed to the reporting of non-significant differences. We at Research the Headlines hope our synopsis and that of NHS Behind the Headlines dampens down the enthusiasm to reject all previous research in this area.

Colen, C. G. & Ramney, D. M. (2014).“Is Breast Truly Best? Estimating the Effects of Breastfeeding on Long-term Child Health and Wellbeing in the United States Using Sibling Comparisons”. Social Science & Medicine, Published online January 29 2014, doi:10.1016/j.socscimed.2014.01.027

From → Biomedical, Health

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: