Breastfeeding: an enduring doctrine
ESSAY: When studies of infant feeding become ‘breast is best’ advocacy, it makes for bad research and bad policy.
Earlier this year, researchers from Essex University’s Institute for Social and Economic Research (ISER) organised an event at the British Academy in London called ‘Early intervention And Social Mobility: Are Pro-Breastfeeding Policies Worth It?’ The findings presented there have become the latest fodder for breastfeeding proponents backing up their advocacy work scientific claims.
At the event, ISER researchers presented results from a set of studies to a panel which included representatives from the Royal College of Midwives, the National Childbirth Trust and Best Beginnings (a charity working ‘to give every baby in the UK the healthiest start in life’). Other panellists included childcare ‘guru’ Dr Miriam Stoppard and Dr Daniel Poulter MP, a member of the House of Commons Health Select Committee.
The event organisers are now planning to convene another meeting with the panellists to agree on ‘three to four key points that we could jointly raise with the chair of the Health Select Committee, Stephen Dorrell’. During a parliamentary question time following the event, Poulter also asked the secretary of state for health, Andrew Lansley, ‘[W]hat steps he is taking to improve long-term support for women on breastfeeding’. Additionally, one of the ISER researchers has presented the findings to the annual conference of the Baby Friendly Initiative (BFI), the influential UNICEF scheme that is shaping procedures in NHS maternity units.
In short, then, the ISER research is becoming the latest example used by some policymakers and pro-breastfeeding lobby groups to claim that ‘the research shows’ that yet more should be done to get women to breastfeed, and that they should do so for longer. Further, this research promises to add a new element to the apparently ‘evidence-based’ case for breastfeeding promotion: the idea that breastfeeding is even better than anyone thought because it can address ‘social mobility’.
Indeed, as the title of the British Academy event suggested, the implication is that breastfeeding might even be best thought of as a kind of ‘early intervention’. Feeding a baby is cast as a sort of policy tool to help make children from ‘socially deprived’ backgrounds much more like their middle-class peers. Or, as the BFI put it, the ISER research shows ‘that breastfeeding may well have a significant part to play in the government’s efforts to reduce inequality’.
What should we make of this example of the relationship between researchers, campaign groups and policymaking? It might look like a model example of what we academics are continually told we are supposed to be doing these days, namely conduct ‘policy-relevant research’ that engages with ‘research users’. As such, one lesson we might learn is that diving head first into this agenda can end up turning academic efforts into something that is in fact quite alien to research, namely ‘advocacy research’.
Indeed, an examination of the details of the ISER research suggests it is hard to draw any other conclusion. In truth, the research tells us very little, if anything at all, about hitherto unrecognised benefits of breastfeeding, but it does say an awful lot about what happens when research morphs into advocacy. The most striking aspect of the academic work and the public account of it is the gap between the public conclusion that breastfeeding can reduce social inequality and the actual findings of the research itself. (A summary of the findings is available here.)
The UK context
Before exploring this gap, it might be useful to clarify what the researchers meant when asking whether or not pro-breastfeeding policies are ‘worth it’. In Britain, policymakers have sought for many years to encourage more women to breastfeed their babies, and to do so for longer. A growing number of government-supported initiatives and interventions have developed to increase breastfeeding rates. Based in maternity wards and in community-based health and social services, such schemes present breastfeeding as unquestionably superior to any other way of feeding a baby. The ISER event revealed that, since 2008, around £6million per year has been spent on breastfeeding promotion, working out at around £7 per child.
This unequivocal emphasis on the importance of breastfeeding, and on changing maternal behaviour to increase the prevalence of the practice, is based on the claim that all the available evidence points to the need for this approach. Policy strongly associates breastfeeding with strategies to address health and social problems. Breastfeeding is promoted as a route to better physical health in mothers and future generations (cancers, coronary heart disease and childhood obesity are highlighted particularly in British policy documents). Social inequality has also been strongly associated by policymakers with breastfeeding for many years, with differential breastfeeding rates by social class put forward to justify the idea there is a relation between infant-feeding practices and larger issues of difference in income and opportunity.
In short, then, policies which promote breastfeeding are thought to be ‘worth it’ if they might save the taxpayer money through reduced costs in the treatment of illness, or by reducing inequality.
Contrary to this apparent consensus, however, recent work in the social sciences has argued that the evidence around the benefits of breastfeeding is much less cut-and-dried than advocates suggest. For instance, the American academic Joan Wolf has disputed theories about the harm of formula feeding. Many are certain that formula feeding is associated with a range of health problems, but in many areas – for example obesity, IQ and psychological development – the evidence is varied and highly inconclusive. Whilst Wolf does demonstrate that breastfeeding protects infants against gastrointestinal infection, on balance the benefits of breastfeeding tend to be dramatically overplayed in the advocacy literature, which routinely conflates correlation with causation, leading to distortion and over-claiming.
At face value, the research at ISER has an admirable objective: trying to find out whether breastfeeding itself causes the benefits to health, cognitive and behavioural outcomes that are usually ascribed to the practice, independently of a mother’s education, IQ, social class and attitudes. To address these questions, four studies were undertaken.
The first study drew on data from the Millennium Cohort Study (MCS) and the UNICEF Baby Friendly Hospital Initiative (BFHI), currently the ‘gold standard’ for care of women in maternity units in official breastfeeding policy. The authors compared rates of breastfeeding in ‘supportive’ settings (ie, BFHI units) with rates in settings that do not offer any special measures (using data from the MCS and other sources). Since mothers do not have any choice over whether they give birth in a BFHI setting, this is, it is argued, as close to a randomised study as possible.
The authors show that the BFHI increases both rates of breastfeeding initiation, and breastfeeding at four weeks of age. However, as to whether this is ‘worth it’ – in terms of reduced costs in treating future illness – the report itself is more sceptical, stating:
‘[T]he causal estimates indicate that breastfeeding exclusively at four weeks reduces the probability that a child will be overweight at three years, but also increases the probability that he will suffer from asthma and wheezing at five and seven years. All other health effects are found to be not significantly different for breastfed vs non-breastfed babies… Most of the effects of breastfeeding on cognitive outcomes are positive, but not statistically significant from zero.’ (Emphasis added.)
The authors’ recommendation to increase investment in the BHFI therefore seems to ignore their own conclusions.
The second study, which examined the effect of breastfeeding on gastrointestinal infection (GI), confirmed what is already known about the effects of breastfeeding in developed countries. As one study has shown, for every 25 breastfed babies in developed countries there would be one fewer GI infection in the first year. As Wolf says, however, ‘That’s a benefit, but it’s fairly minimal, and parents need to decide whether the benefit is worth the cost.’
What’s troubling, then, is the way in which these minimal differences are presented to parents: in the ISER report about this study it is shown that 18 in every 10,000 babies (0.0018 per cent) who were not breastfed were hospitalised in the first year with a GI infection compared with five in every 10,000 babies (0.0005 per cent) who were breastfed.
These are, clearly, minuscule proportions and they could help communicate a very reassuring message to mothers that however you feed your baby, it is very unlikely that it will end up in hospital with a serious tummy bug. Instead, the statistics are represented as a warning of the dangers of formula feeding: ‘Exclusively breastfed babies are 2.7 times less likely to be hospitalised’ than formula-fed babies, states the report.
The second section of this report, which addresses how long these protective aspects of breastfeeding last is also problematic. It fails to take account of the fact that it is possible that those children who are formula fed are more likely to be in collective settings (such as nurseries) than those who remain exclusively breastfed, or to spend more time in contact with other children for other reasons. Yet such factors, which may account for why they are likely to catch infections and which might affect results showing there is a ‘doubling of the risk of being hospitalised for diarrhoea for every month since stopping breastfeeding’, are not commented on. This is a flaw in research looking into the benefits of breastfeeding that was identified 20 years ago, and it is therefore very striking that the authors of this study failed to give space to its discussion.
The third study (which partly gives rise to the claim that ‘government should consider breastfeeding as a social mobility measure’) examined breastfeeding and what the authors term ‘socio-emotional development’, based on a child’s behaviour at five years old. Again, using data from the MCS (whereby parents were interviewed when their child was nine months old and then revisited at two-yearly intervals) parents were asked to complete a ‘strengths and difficulties questionnaire’. Here, they were asked to score their children and identify any potential behavioural problems, including emotional problems (such as clinginess and anxiety), hyperactivity (restlessness), and conduct problems (lying and stealing). The report argues that ‘in full-term children, longer duration of breastfeeding (four months or more) is associated with fewer parent-rated behavioural problems at age five’.
Questions might be raised at the outset about the categorisation of the behaviours listed above as ‘problems’ and the certainty with which they might be considered to have long-term implications or a causal connection to ‘social mobility’. This is not the only reason, however, why it could be considered unwise to generate policy proposals on the basis of this research.
A further problem is that, like the second study, it relies on the self-reporting of parents, a methodology which has been shown to be highly unreliable. This is especially the case in a context of ‘intensive motherhood’ where a strong performative element of mothering needs to be recognised. That is, it is plausible that mothers who formula feed and mothers who exclusively breastfeed are not equally likely to report that their children have been ill or have behavioural problems, particularly if they have internalised the injunctions around breastfeeding being best. Breastfeeding mothers may have a greater investment in their children’s development which could cause them to ‘see’ fewer problems (and therefore affirm that they are good parents).
Lastly, this research is unable to untangle whether the effects observed are down to breastmilk (and the potential effect of the long-chain polyunsaturated fatty acids it contains) or the practice of breastfeeding itself (which is said to create greater interaction between the mother and child). It is well known that breastfed babies do tend to be healthier than formula fed babies on average, but this may be more to do with other behaviours surrounding breastfeeding than breastfeeding per se.
The problem with this sort of research is that even when various factors such as class, income and education are controlled for, one can never control for a woman’s characteristics. A mother’s decision to breastfeed represents more than her social class; it reflects a certain orientation towards parenting that itself might be correlated with certain outcomes in terms of IQ or health.
There appears to be an assumption that if benefits associated with breastfeeding cannot be explained by class, they must be the result of breastfeeding. Yet the type of mother who is likely to internalise the message about the benefits of breastfeeding is also more likely to internalise other sorts of advice around healthy eating, or ‘good parenting’. In short, one can never control for a particular woman’s decision to breastfeed. Research demonstrating that benefits cannot be explained by class does not demonstrate that they are necessarily explained by breastfeeding.
The final study explored the relationship between breastfeeding and intelligence – historically one of the most controversial areas in these debates. Intelligence was judged on children’s test scores in reading, writing and mathematics at ages five, seven, 11 and 14. The study used a technique called twinning. This is where each breastfed child from a data set (the Avon Longitudinal Study of Parents and Children, ALSPAC) was twinned with one or more children who were not breastfed and from a similar demographic background. The researchers attempted to identify the ‘true’ effect of breastfeeding on children’s cognitive development. Importantly, they say, the babies are ‘twinned’ on the parents’ attitudes, and notably this includes their intentions to breastfeed or not.
Again, this study fails to take into account that one can never control for a mother’s decision to breastfeed, particularly in a highly moralised context like that of infant feeding, where even the intentions one reports act as a barometer of social orthodoxies. The researchers state that the effect of breastfeeding is not large, but this does not prevent them from also stating that ‘children breastfed for four weeks or more do better than children breastfed for less than four weeks by about one tenth of a standard deviation… [T]his loosely translates to a difference of about three IQ points.’ In fact, one could also argue that the more carefully the behavioural environment is controlled for, the less that the method of infant feeding seems to matter.
A final point that relates to all four studies is that where research sets out to find a correlation between certain behaviours and particular outcomes, this in itself reflects a prior assumption of causation. It is unlikely, for example, that a research study would set out to find a link between formula feeding and intelligence.
What’s the problem with all this?
The most notable feature of this research is the gulf between the data and the public account of those data. The narrative that has emerged is one about the benefits of breastfeeding over formula feeding to the extent of suggesting that breastfeeding can help increase the social mobility of working-class children. It is argued that increased funds ought to be invested in these initiatives. For example, a recent article in the Observer was titled ‘Breastfed babies are smarter and get a better chance in life’ and explained that ‘The government is urged to act on new research that shows the importance of a baby’s early weeks to success in later life’. Yet the data actually seem quite ambiguous and open to a variety of interpretations.
In this light, the ISER research, and the advocacy around breastfeeding that it is bound up with, reflects the wider realities around infant feeding. That is, there is a gap between the one-sided and inflexible policy perspective that insists that there is only one good way to feed a baby, and the far more ambivalent maternal experiences. As other research in this field has shown, these one-sided accounts create unnecessary anxiety in mothers, a problem that should be taken far more seriously by policymakers and advocates. There are many reasons that women might consider breastfeeding ‘worth it’ – because their children enjoy it, because they find it more convenient or any number of other reasons – but this didacticism around health outcomes is disingenuous. Certainly, as Wolf’s analysis makes clear, a policy discourse which attaches benefits to breastfeeding and risk to formula feeding demeans the importance of informed choice, by failing to convey actual evidence around infant feeding to mothers.
No one will argue with women being given support in carrying out their infant-feeding decisions. But there is a difference between support and advocacy which pro-breastfeeding policies elide. It is important to recognise that the reason so many women want to breastfeed is the product of a larger culture which ties breastfeeding to doing what is best for one’s child. It is seen as part of ‘good motherhood.’ This is itself a cultural construction – or, as Eric Hobsbawm might put it, an invented tradition at a specific historical juncture. Policy shapes these very desires, as much as the other way around.
The ISER reports conclude by saying that there should be more support for pro-breastfeeding policies, such as the UNICEF Baby Friendly Initiative. It claims to have revealed that ‘the duration of breastfeeding…might be crucial to achieving the long-term benefits in health, cognitive and behavioural development. Encouraging more women to initiate breastfeeding is the starting point, but ensuring that the support is there in the form of health visitors and peer support groups so that more women continue to breastfeed exclusively for longer has to be a policy goal.’
In fact, the report findings could easily have been used to make a very different argument. In moving away from the effort to be objective, and becoming advocacy research instead, this research has both lost opportunities and created problems. It has encouraged a misleading idea about what research is and should strive to be; an open-ended, provisional account rightfully open to interpretation and rightfully subject to genuine debate. In opting to set aside detachment for advocacy work, the ISER reports have added to the wealth of confusion and anxiety around ways of feeding babies.
Ellie Lee and Charlotte Faircloth work in the Centre for Parenting Culture Studies at the University of Kent. Ellie Lee is the author of a research briefing that looks further at the issues raised in this article.
To enquire about republishing spiked’s content, a right to reply or to request a correction, please contact the managing editor, Viv Regan.
Want to join the conversation?
Only spiked supporters, who donate regularly to us, can comment on our articles.